Modelling the persistence of infectious diseases in pre-agricultural Hunter-gatherers

Final Analysis Report

Set-up

rm(list = ls())
library(dplyr)
library(wesanderson)
library(GillespieSSA)
library(tidyverse)

Background

The traditionally held belief that most modern infectious diseases emerged when humans began living in larger agricultural settlements has been challenged by studies of modern hunter-gatherers. This study will investigate which emerging pathogens may persist in hunter-gatherer groups by constructing a compartment model of infectious disease transmission that accounts for demography and multi-band structure. This study will look to understand how the critical community size required to sustain an outbreak is affected by host population dynamics. We show that metapopulation structure increases the probability of a respiratory pathogen with waning immunity persisting after 3 years. The probability of persistence increases with the number of sub-populations but is largely determined by the duration of immunity. Understanding the origins of infectious diseases is an important area of research that will lead to improved strategies for reducing their global burden.

This report will cover the full analysis undertaken to generate the results used in my MSc project. A full description of the research project aims and methods can be found in the final paper in the Hunter_Gatherer_models GitHub repository. Some code used in this project was adapted from the tutorials attached to the GillespieSSA package.

Model Parameter estimation

Agta Hunter-Gatherer Demography

Modern-day hunter-gatherers are often used to make inferences about pre-agricultural human populations. This study modeled the host population on a group of indigenous hunter-gatherers from the Northern Phillipines known as the Agta. Information regarding births, deaths and population size were obtain from a study conducted by Headland et al., (2011). Authors conducted a census-like survey of the Agta that followed $$4,300 individuals over the period of 1950-2010. This date was first explored to understand Agta demography.

agta_demo <- read.csv("Agta_Data/AgtaPopDynamics_Headland2007.csv")

ggplot(agta_demo, aes(x=Year)) +
  geom_line(aes(y=PopSize), colour = wes_palettes$Darjeeling1[1]) +
  geom_line(aes(y=Births), colour = wes_palettes$Darjeeling1[2]) +
  geom_line(aes(y=Deaths), colour = wes_palettes$Darjeeling1[3]) +
  theme_bw()

Population Size

hist(agta_demo$PopSize)

summary(agta_demo$PopSize)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
  133.0   177.0   213.0   211.4   228.0   295.0 

Births

hist(agta_demo$Births)

summary(agta_demo$Births)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
   4.00    8.00   10.00   10.33   12.25   15.00       1 

Deaths

hist(agta_demo$Deaths)

summary(agta_demo$Deaths)

   Min. 1st Qu.  Median    Mean 3rd Qu.    Max.    NA's 
  2.000   5.000   7.000   7.683  10.000  23.000       1 

Birth/Death rate per person per day

Birth rate was estimated from this data by taking the mean of the annual number of births divided by two times the annual number of females. This was then scaled appropriately to obtain the daily mean birth rate per person.

agta_demo <- agta_demo %>%
  mutate(Birth_rate = Births/(Female*2),
         Birth_rate_daily = (1 + Birth_rate) ^ (1/365) - 1,
         Death_rate = (Deaths/PopSize),
         Death_rate_daily = (1 + Death_rate) ^ (1/365) - 1,
         PopChange = (diff = PopSize - lag(PopSize, default = first(PopSize))),
         PopChange_rate = abs(PopChange)/PopSize,
         PopChange_rate_daily = (1 + PopChange_rate) ^ (1/365) - 1)
head(agta_demo)

hist(agta_demo$Birth_rate)

hist(agta_demo$Death_rate)

ggplot(agta_demo, aes(x=Year)) +
  geom_line(aes(y=Birth_rate), colour = wes_palettes$Darjeeling1[2]) +
  geom_line(aes(y=Death_rate), colour = wes_palettes$Darjeeling1[3]) +
  theme_bw()

demo_sum <- agta_demo %>%
  select(PopSize, Birth_rate, Birth_rate_daily, Death_rate, Death_rate_daily, PopChange_rate, PopChange_rate_daily) %>%
    summarise(across(
    .cols = is.numeric, 
    .fns = list(Mean = mean, SD = sd), na.rm = TRUE, 
    .names = "{col}_{fn}"
    ))
demo_sum 

demo_sum <- as.list(demo_sum)

Agta Band Size

Data regarding camp size of Agta hunter-gatherers was obtained from a study of 615 individuals from 15 camps in in the municipality of Palanan, the Northern Philippines published by Dyble et al. (2021).

# Import Camp data from Mark Dyble
camps.data <- read_csv("Agta_Data/camps.csv")

New names:Rows: 15 Columns: 9── Column specification ──────────────────────────────────────────────────────
Delimiter: ","
chr (1): camp_name
dbl (8): ...1, camp_total, camp_adult_men, camp_adult_women, camp_all_r, c...
ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

head(camps.data)

# Explore camp size
hist(camps.data$camp_total)

camp.size <- camps.data %>%
  summarise(mean = mean(camp_total),
            sd = sd(camp_total),
            min = min(camp_total),
            max = max(camp_total),
            var = var(camp_total))
camp.size

Pathogen X

For the purpose of this investigation we formulated a hypothetical respiratory pathogen, referred to as pathogen X. Taking into account the biological trade-offs between high transmissibility and high pathogenicity, pathogen X was decided to be highly infectious with a relatively low case fatality rate of 0.005. Transmission occurred via close contact with an infected individual. Infection was characterised by a latent period of 5.7 days followed by an infectious period of 5 days. Individuals who recovered from infection were immune for 100 days, after which immunity waned and individuals became susceptible to re-infection. Based on these characteristics, the parameters in table 1 were assumed and input into the final models.

Parameter	Rate	Value
\(\beta\)	Transmission	0.6
\(\sigma\)	Infectious	0.175
\(\gamma\)	Recovery	0.2
\(\alpha\)	Death from Infection	0.001
\(\omega\)	Waning Immunity	0.01

Single Population Model

To investigate the persistence of a hypothetical respiratory pathogen in hunter-gatherers, this study chose to simulate disease transmission using a compartment model approach as outlined in the introduction. Two models were constructed to investigate compare the effect of metapopulation structure on disease persistence. This first describes the transmission of a pathogen within a single population with demography and waning immunity to re-infection over time.

Figure 1 - Flow diagram of SEIRS model of transmission

\[\begin{align*} \frac{{{\mathrm{d}}S}}{{{\mathrm{d}}t}} & = \underbrace {\mu N}_{{\mathrm{birth}}}~ - ~\underbrace {\frac{\beta SI}{N}}_{{\mathrm{infection}}}~~ + \underbrace {\omega R}_{{\mathrm{lost}}\,{\mathrm{immunity}}} - \underbrace {\mu S}_{{\mathrm{death}}} \\ \frac{{{\mathrm{d}}E}}{{{\mathrm{d}}t}} & = \underbrace {\frac{\beta SI}{N}}_{{\mathrm{infection}}}~ - ~\underbrace {\sigma E}_{{\mathrm{latency}}} - \underbrace {\mu E}_{{\mathrm{death}}} \\ \frac{{{\mathrm{d}}I}}{{{\mathrm{d}}t}} & = \underbrace {\sigma E}_{{\mathrm{latency}}} - \underbrace {\gamma I}_{{\mathrm{recovery}}} - ~\underbrace {\left( {\mu + \alpha } \right)I}_{{\mathrm{death}}} \\ \frac{{{\mathrm{d}}R}}{{{\mathrm{d}}t}} & = \underbrace {\gamma I}_{{\mathrm{recovery}}} - \underbrace {\omega R}_{{\mathrm{lost}}\ {\mathrm{immunity}}} - \underbrace {\mu R}_{{\mathrm{death}}} \end{align*}\]

Where transmission is frequency dependent, \({\frac{\beta SI}{N}}\), \(\frac{1}{\sigma}\) is the duration of the latent phase, \(\frac{1}{\gamma}\) is the duration of infection, \(\frac{1}{\omega}\) is the duration of immunity and death from infection occurs at the rate \(\alpha\). Individuals can be born into S and die naturally from any compartment at a rate of \(\mu\).

Model Set-up

Model was set up with a single randomly selected camp size with a single infected individual and parameters for pathogen X.

# Define Paramenters
N <-    sample(camps.data$camp_total, 1)    # Population size
initial_infected <-  1    # Initial infected
simName <- "SEIRS model"       # Simulation name
tf <- 365*3

#Collect parameters
parms <- list(
  beta = 0.6,
  sigma = 0.175,                          # E to I rate
  gamma = 0.2,                           # I to R rate
  omega = 1/100,                         # R to S rate
  mu = demo_sum$Birth_rate_daily_Mean,                            # Birth/death rate per person per day
  alpha = 1/1000) 

#Create the named initial state vector for the U-patch system.

x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

names(x0) <- c("S","E","I", "R", "N")


# Define the state change matrix for a single patch
nu <- matrix(c( -1,  0,  0, +1, +1, -1,  0,  0,  0,  0, # S
                +1, -1,  0,  0,  0,  0, -1,  0,  0,  0, # E
                0, +1, -1,  0,  0,  0,  0, -1,  0, -1, # I
                0,  0, +1, -1,  0,  0,  0,  0, -1,  0, # R 
                0,  0,  0,  0, +1, -1 ,-1, -1, -1, -1), # N
             nrow=5,byrow=TRUE)

# Define propensity functions
a <-c(
        paste0("(beta*I/N)*S"), # Infection
        paste0("sigma*E"),                                       # Becomes infecious
        paste0("gamma*I"),                                       # Recovery from infection
        paste0("omega*R"),       # Loss of immunity
        paste0("mu*N"),                             # Births
        paste0("mu*S"),                                             # Deaths (S)
        paste0("mu*E"),                                             # Deaths (E)
        paste0("mu*I"),                                             # Deaths (I)
        paste0("mu*R"),                                             # Deaths (R)
        paste0("alpha*I")                                           # Deaths from infection
        
      )

Define functions to calculate R0 and expected number of susceptibles at equilibrium, and critical community size (Diekmann et al., 2012).

 R0 <- function(parms) {
   (parms$sigma/(parms$sigma + parms$mu)) * (parms$beta/parms$gamma + parms$mu + parms$alpha)
 } 
  
EIE <- function(R0, parms) {
  y = ((R0 - 1) * parms$omega) / (parms$gamma * R0)
  return(y)
}

CCS <- function(epsilon, R0) {
  y = 1/((epsilon^2)*((1-(1/R0))^2))
  return(y)
}

Run Single Population Model

# Calculate R0, expected number of infecteds at equilibrium, magnitude of oscillation and CCS
R0_single <- R0(parms)
R0_single

EIE_single <- EIE(R0_single, parms) # proportion of expected infecteds at equilibrium
EIE_single

expexted_infecteds <- EIE_single*N # number of expected infecteds at equilibrium
expexted_infecteds

sqrt(N) # magnitude of oscillations 

epsilon <- (5.7/365)/23 # duration of infection in years divided by avg life expectancy 
CCS_single <- CCS(epsilon, R0_single) # Average life expectancy as per Kaplan (crude)
CCS_single

# Run simulations with the Direct method
set.seed(21)
out <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
) 



## Extra Plots
plot_data <- out$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

single_plot <- ggplot(data = plot_data, aes(x=t, y=count, colour=state))+
  geom_line(alpha=0.8)+
  labs(x="Time (Days)",
       y="Number of Individuals", 
       colour="State")+
  geom_hline(yintercept = expexted_infecteds, linetype = 'dashed') +
  theme_bw()

single_plot

ggsave(filename = "single_plot.pdf", 
       plot = single_plot,
       device = "pdf",
       width = 7, 
       height = 3,
       path = "/Users/matthewhoyle/Github_R_projects/Plots/Hunter_Gatherer_models")

plot_data %>%
  filter(state == "I") %>%
  slice_max(count)

Outbreak peaked at day 25 with 14 infected individuals.

## Run multiple simulations and saving output
num_sims <- 1000
sim_list <- list()
sim_list <- vector("list", length = num_sims)

for (i in 1:num_sims){
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  set.seed(i)
  out_100 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data <- out_100$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list[[i]] <- sim_data
}

sim_output <- bind_rows(sim_list)

# Summary table of endpoint data
sim_output <- sim_output %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
head(sim_output)

# Make Summary Table of output
sim_summary <- sim_output %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100, 
            mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/100)
sim_summary

Varying waining immunity

Waning immunity was thought to play an important role in the persistence of pathogen X so we incrementally increased the duration of immunity (by decreasing \(\omega\)) and calculated the probability of persistence after 3 years in 1000 stochastic simulations. Duration of immunity was increased from 1 day to a year.

0 Days

#Collect parameters
parms_0 <- parms
parms_0$omega <- 0


# Run simulations with the Direct method
set.seed(4)
out_0 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_0,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_0 <- out_0$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_0 <- ggplot(data = plot_data_0, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_0

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_0 <- list()
sim_list_0 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_0 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_0,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_0 <- out_100_0$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_0[[i]] <- sim_data_0
}

sim_output_0 <- bind_rows(sim_list_0)

# Summary table of endpoint data
sim_output_0 <- sim_output_0 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_0

# Make Summary Table of output
sim_summary_0 <- sim_output_0 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 0)
sim_summary_0

1 Days

#Collect parameters
parms_1 <- parms
parms_1$omega <- 1


# Run simulations with the Direct method
set.seed(4)
out_1 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_1,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_1 <- out_1$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_1 <- ggplot(data = plot_data_1, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_1

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_1 <- list()
sim_list_1 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_1 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_1,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_1 <- out_100_1$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_1[[i]] <- sim_data_1
}

sim_output_1 <- bind_rows(sim_list_1)

# Summary table of endpoint data
sim_output_1 <- sim_output_1 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_1

# Make Summary Table of output
sim_summary_1 <- sim_output_1 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1)
sim_summary_1

3 Days

#Collect parameters
parms_3 <- parms
parms_3$omega <- 1/3


# Run simulations with the Direct method
set.seed(4)
out_3 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_3,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3 <- out_3$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3 <- ggplot(data = plot_data_3, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3 <- list()
sim_list_3 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_3 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_3,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3 <- out_100_3$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_3[[i]] <- sim_data_3
}

sim_output_3 <- bind_rows(sim_list_3)

# Summary table of endpoint data
sim_output_3 <- sim_output_3 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3

# Make Summary Table of output
sim_summary_3 <- sim_output_3 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/3)
sim_summary_3

7 Days

#Collect parameters
parms_7 <- parms
parms_7$omega <- 1/7


# Run simulations with the Direct method
set.seed(4)
out_7 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_7,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_7 <- out_7$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_7 <- ggplot(data = plot_data_7, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_7

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_7 <- list()
sim_list_7 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_7 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_7,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_7 <- out_100_7$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_7[[i]] <- sim_data_7
}

sim_output_7 <- bind_rows(sim_list_7)

# Summary table of endpoint data
sim_output_7 <- sim_output_7 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_7

# Make Summary Table of output
sim_summary_7 <- sim_output_7 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/7)
sim_summary_7

10 Days

#Collect parameters
parms_10 <- parms
parms_10$omega <- 1/10


# Run simulations with the Direct method
set.seed(4)
out_10 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_10,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_10 <- out_10$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_10 <- ggplot(data = plot_data_10, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_10

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_10 <- list()
sim_list_10 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_10 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_10,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_10 <- out_100_10$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_10[[i]] <- sim_data_10
}

sim_output_10 <- bind_rows(sim_list_10)

# Summary table of endpoint data
sim_output_10 <- sim_output_10 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_10

# Make Summary Table of output
sim_summary_10 <- sim_output_10 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/10)
sim_summary_10

20 Days

#Collect parameters
parms_20 <- parms
parms_20$omega <- 1/20


# Run simulations with the Direct method
set.seed(4)
out_20 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_20,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_20 <- out_20$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_20 <- ggplot(data = plot_data_20, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_20

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_20 <- list()
sim_list_20 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_20 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_20,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_20 <- out_100_20$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_20[[i]] <- sim_data_20
}

sim_output_20 <- bind_rows(sim_list_20)

# Summary table of endpoint data
sim_output_20 <- sim_output_20 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_20

# Make Summary Table of output
sim_summary_20 <- sim_output_20 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/20)
sim_summary_20

30 Days

#Collect parameters
parms_30 <- parms
parms_30$omega <- 1/30


# Run simulations with the Direct method
set.seed(4)
out_30 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_30,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_30 <- out_30$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_30 <- ggplot(data = plot_data_30, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_30

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_30 <- list()
sim_list_30 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_30 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_30,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_30 <- out_100_30$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_30[[i]] <- sim_data_30
}

sim_output_30 <- bind_rows(sim_list_30)

# Summary table of endpoint data
sim_output_30 <- sim_output_30 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_30

# Make Summary Table of output
sim_summary_30 <- sim_output_30 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/30)
sim_summary_30

40 Days

#Collect parameters
parms_40 <- parms
parms_40$omega <- 1/40


# Run simulations with the Direct method
set.seed(4)
out_40 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_40,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_40 <- out_40$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_40 <- ggplot(data = plot_data_40, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_40

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_40 <- list()
sim_list_40 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_40 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_40,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_40 <- out_100_40$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_40[[i]] <- sim_data_40
}

sim_output_40 <- bind_rows(sim_list_40)

# Summary table of endpoint data
sim_output_40 <- sim_output_40 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_40

# Make Summary Table of output
sim_summary_40 <- sim_output_40 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/40)
sim_summary_40

50 Days

#Collect parameters
parms_50 <- parms
parms_50$omega <- 1/50


# Run simulations with the Direct method
set.seed(4)
out_50 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_50,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_50 <- out_50$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_50 <- ggplot(data = plot_data_50, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_50

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_50 <- list()
sim_list_50 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_50 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_50,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_50 <- out_100_50$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_50[[i]] <- sim_data_50
}

sim_output_50 <- bind_rows(sim_list_50)

# Summary table of endpoint data
sim_output_50 <- sim_output_50 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_50

# Make Summary Table of output
sim_summary_50 <- sim_output_50 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/50)
sim_summary_50

60 Days

#Collect parameters
parms_60 <- parms
parms_60$omega <- 1/60


# Run simulations with the Direct method
set.seed(4)
out_60 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_60,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_60 <- out_60$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_60 <- ggplot(data = plot_data_60, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_60

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_60 <- list()
sim_list_60 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_60 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_60,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_60 <- out_100_60$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_60[[i]] <- sim_data_60
}

sim_output_60 <- bind_rows(sim_list_60)

# Summary table of endpoint data
sim_output_60 <- sim_output_60 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_60

# Make Summary Table of output
sim_summary_60 <- sim_output_60 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/60)
sim_summary_60

70 Days

#Collect parameters
parms_70 <- parms
parms_70$omega <- 1/70


# Run simulations with the Direct method
set.seed(4)
out_70 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_70,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_70 <- out_70$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_70 <- ggplot(data = plot_data_70, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_70

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_70 <- list()
sim_list_70 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_70 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_70,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_70 <- out_100_70$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_70[[i]] <- sim_data_70
}

sim_output_70 <- bind_rows(sim_list_70)

# Summary table of endpoint data
sim_output_70 <- sim_output_70 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_70

# Make Summary Table of output
sim_summary_70 <- sim_output_70 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist)/num_sims)*100,
            mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/70)
sim_summary_70

80 Days

#Collect parameters
parms_80 <- parms
parms_80$omega <- 1/80


# Run simulations with the Direct method
set.seed(4)
out_80 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_80,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_80 <- out_80$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_80 <- ggplot(data = plot_data_80, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_80

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_80 <- list()
sim_list_80 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_80 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_80,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_80 <- out_100_80$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_80[[i]] <- sim_data_80
}

sim_output_80 <- bind_rows(sim_list_80)

# Summary table of endpoint data
sim_output_80 <- sim_output_80 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_80

# Make Summary Table of output
sim_summary_80 <- sim_output_80 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/80)
sim_summary_80

150 Days

#Collect parameters
parms_150 <- parms
parms_150$omega <- 1/150


# Run simulations with the Direct method
set.seed(4)
out_150 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_150,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_150 <- out_150$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_150 <- ggplot(data = plot_data_150, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_150

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_150 <- list()
sim_list_150 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_150 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_150,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_150 <- out_100_150$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_150[[i]] <- sim_data_150
}

sim_output_150 <- bind_rows(sim_list_150)

# Summary table of endpoint data
sim_output_150 <- sim_output_150 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_150

# Make Summary Table of output
sim_summary_150 <- sim_output_150 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/150)
sim_summary_150

180 Days

#Collect parameters
parms_180 <- parms
parms_180$omega <- 1/180


# Run simulations with the Direct method
set.seed(4)
out_100 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_180,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_180 <- out_180$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_180 <- ggplot(data = plot_data_180, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_180

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_180 <- list()
sim_list_180 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_180 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_180,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_180 <- out_100_180$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_180[[i]] <- sim_data_180
}

sim_output_180 <- bind_rows(sim_list_180)

# Summary table of endpoint data
sim_output_180 <- sim_output_180 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_180

# Make Summary Table of output
sim_summary_180 <- sim_output_180 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/180)
sim_summary_180

365 Days

#Collect parameters
parms_365 <- parms
parms_365$omega <- 1/365


# Run simulations with the Direct method
set.seed(4)
out_365 <- ssa(
  x0 = x0,
  a = a,
  nu = nu,
  parms = parms_365,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_365 <- out_365$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "state", values_to = "count") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_365 <- ggplot(data = plot_data_365, aes(x=t, y=count, colour=state))+
  geom_line()+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_365

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_365 <- list()
sim_list_365 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  N <-     sample(camps.data$camp_total, 1)    # Sample different patch sizes for each sim
  
  x0 <- c(N - initial_infected, initial_infected, 0, 0, N)

  names(x0) <- c("S","E","I", "R", "N")


  out_100_365 <- ssa(
    x0 = x0,
    a = a,
    nu = nu,
    parms = parms_365,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_365 <- out_100_365$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "state", values_to = "count") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, I, N, persist)
  
  sim_list_365[[i]] <- sim_data_365
}

sim_output_365 <- bind_rows(sim_list_365)

# Summary table of endpoint data
sim_output_365 <- sim_output_365 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_365

# Make Summary Table of output
sim_summary_365 <- sim_output_365 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/365)
sim_summary_365

Results

waning_results_single <- sim_summary %>%
  bind_rows(sim_summary_1) %>%
  bind_rows(sim_summary_3) %>%
  bind_rows(sim_summary_7) %>%
  bind_rows(sim_summary_10) %>%
  bind_rows(sim_summary_20) %>%
  bind_rows(sim_summary_30) %>%
  bind_rows(sim_summary_40) %>%
  bind_rows(sim_summary_50) %>%
  bind_rows(sim_summary_60) %>%
  bind_rows(sim_summary_70) %>%
  bind_rows(sim_summary_80) %>%
  bind_rows(sim_summary_100) %>%
  bind_rows(sim_summary_150) %>%
  bind_rows(sim_summary_365) %>%
  mutate(immunity_duration = 1/omega) %>%
  arrange(immunity_duration) %>%
  mutate(model="single",
         patches = 1)

write_csv(waning_results_single, file = "/Users/matthewhoyle/Github_R_projects/Hunter_Gatherer_models/Results/waning_results_single.csv")

waning_results_single

ggplot(waning_results_single, aes(immunity_duration, sum_persist)) +
  geom_line()+
  geom_point()+
  theme_bw()

7-Patch Metapopulation Model

Recent studies have suggested that pre-agricultural hunter-gatherers did not live in small isolated groups but fromed interconnected multi-camp networks. To investigate this we built a second population that accounts for metapopulation structure of hunter-gatherers. The second model follows an almost identical format as the single population model, but instead has been expanded to accommodate the metapopulation structure of multi-band hunter-gatherer groups:

\[\begin{align*} \frac{{{\mathrm{d}}S}}{{{\mathrm{d}}t}} & = \underbrace {\mu_i N_i}_{{\mathrm{birth}}}~ - ~\underbrace {\biggl(\frac{\beta_{ii} I_i}{N_i} + \frac{\beta_{ji} I_j} {N_j} + ... \biggr)S_i}_{{\mathrm{infection}}}~~ + \underbrace {\omega_i R_i}_{{\mathrm{lost}}\,{\mathrm{immunity}}} - \underbrace {\mu_i S_i}_{{\mathrm{death}}} \\ \frac{{{\mathrm{d}}E}}{{{\mathrm{d}}t}} & = \underbrace {\biggl(\frac{\beta_{ii} I_i}{N_i} + \frac{\beta_{ji} I_j} {N_j} + ... \biggr)S_i}_{{\mathrm{infection}}}~ - ~\underbrace {\sigma_i E_i}_{{\mathrm{latency}}} - \underbrace {\mu_i E_i}_{{\mathrm{death}}} \\ \frac{{{\mathrm{d}}I}}{{{\mathrm{d}}t}} & = \underbrace {\sigma_i E_i}_{{\mathrm{latency}}} - \underbrace {\gamma_i I_i}_{{\mathrm{recovery}}} - ~\underbrace {\left( {\mu_i + \alpha_i } \right)I_i}_{{\mathrm{death}}} \\ \frac{{{\mathrm{d}}R}}{{{\mathrm{d}}t}} & = \underbrace {\gamma_i I_i}_{{\mathrm{recovery}}} - \underbrace {\omega_i R_i}_{{\mathrm{lost}}\ {\mathrm{immunity}}} - \underbrace {\mu_i R_i}_{{\mathrm{death}}} \end{align*}\]

These coupled differential equations describe the within-patch SEIRS-type dynamics of the \(i\)th patch where the force of infection is driven by contact of susceptibles with infecteds within the \(i\)th patch and in the \(j\)th other patches. Both models assume that compartments are well-mixed and that the waiting times between compartments are exponentially distributed.

Model Set-up

We first modeled transmission in a metapopulation of 7 camps, as observed by Migliano et al. (2023), with one initially infected individual from a randomly selected patch.

# Define Paramenters
patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Patch size
U <- length(patchPopSize)                    # Number of patches
initial_infected <-  as.vector(rmultinom(1, 1, rep(0.5, U)))   # Initial infected (initial infected patch randomly generated)
initial_infected_patch <- which(initial_infected > 0)
simName <- "SIRS metapopulation model"       # Simulation name
tf <- 365*3                                   # Final time

# Agta Hunter-Gatherer contact rates
within_pop_contact = 1
between_pop_contact = 0.5/U     # normalised by number of patches 

#Create the named initial state vector for the U-patch system.

x0_meta <- unlist(lapply(
  seq_len(U), 
  function(i){ 
    c(patchPopSize[i] - initial_infected[i], initial_infected[i], 0, 0, patchPopSize[i])
  }
))

names(x0_meta) <- unlist(lapply(seq_len(U), function(i) paste0(c("S","E","I", "R", "N"), i)))


# Define the state change matrix for a single patch
nu_meta <- matrix(c( -1,  0,  0, +1, +1, -1,  0,  0,  0,  0, # S
                     +1, -1,  0,  0,  0,  0, -1,  0,  0,  0, # E
                      0, +1, -1,  0,  0,  0,  0, -1,  0, -1, # I
                      0,  0, +1, -1,  0,  0,  0,  0, -1,  0, # R 
                      0,  0,  0,  0, +1, -1 ,-1, -1, -1, -1), # N
             nrow=5,byrow=TRUE)

# Define propensity functions
# Mass-action
a_meta <-
  unlist(lapply(
    seq_len(U),
    function(patch) {
      i <- patch
      patches <- 1:U
      #j <- if (patch == 1) U else patch - 1
      other_patches <- patches[-i]
      patch_beta <- c()
      for(k in (1:(U-1))){
        patch_beta[k] = paste0("+(beta_", other_patches[k],i, "*I", other_patches[k], "/N", other_patches[k], ")*S", i)
      }
      c(
        paste0("(beta_", i, i, "*I", i,"/N", i, ")*S",i, paste0(patch_beta, collapse="")), # Infection
        paste0("sigma*E", i),                                       # Becomes infecious
        paste0("gamma*I", i),                                       # Recovery from infection
        paste0("omega*R", i),       # Loss of immunity
        paste0("mu*N", i),                             # Births
        paste0("mu*S", i),                                             # Deaths (S)
        paste0("mu*E", i),                                             # Deaths (E)
        paste0("mu*I", i),                                             # Deaths (I)
        paste0("mu*R", i),                                             # Deaths (R)
        paste0("alpha*I", i)                                           # Deaths from infection
        
      )
    }
  ))

Define functions for calculating R0 from next-generation matrix

# Calculate R0 from NGM

R0ngm <- function(nextgen_matrix) {
  eigenvalues = eigen(nextgen_matrix, only.values = T)
  R0 = max(abs(eigenvalues$values))
  return(R0)
}

beta.ngm <- function(beta_matrix) {
  eigenvalues = eigen(beta_matrix, only.values = T)
  beta_ngm = max(abs(eigenvalues$values))
  return(beta_ngm)
}

Run Metapopulation Model

#Collect parameters
parms_meta <- list(
  sigma = 0.175,                          # E to I rate
  gamma = 0.2,                           # I to R rate
  omega = 1/100,                         # R to S rate
  mu = demo_sum$Birth_rate_daily_Mean,                            # Birth/death rate per person per day
  alpha = 1/1000) 

# Define transmission terms and populate next-generation matrix
beta <- 0.6

nextgen_matrix <- matrix(nrow = U, ncol = U, data = 0)
beta_matrix <- matrix(nrow = U, ncol = U, data = 0)


for(i in 1:U){
  for(j in 1:U){
    parms_meta[[paste0("beta_",i,i)]] = within_pop_contact*beta
    nextgen_matrix[i,i] = within_pop_contact*beta*(1/parms_meta$gamma)
    parms_meta[[paste0("beta_",j,i)]] = between_pop_contact*beta
    nextgen_matrix[j,i] = between_pop_contact*beta*(1/parms_meta$gamma)
    nextgen_matrix[i,j] = between_pop_contact*beta*(1/parms_meta$gamma)
    parms_meta[[paste0("beta_",j,j)]] = within_pop_contact*beta
    nextgen_matrix[j,j] = within_pop_contact*beta*(1/parms_meta$gamma)
    beta_matrix[i,i] = within_pop_contact*beta
    beta_matrix[j,i] = between_pop_contact*beta
    beta_matrix[i,j] = between_pop_contact*beta
    beta_matrix[j,j] = within_pop_contact*beta
  }
  parms_meta[[paste0("N", i)]] = patchPopSize[i]
}

# Run simulations with the Direct method
set.seed(25)
out_meta <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Plot
plot_data_meta <- out_meta$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta <- ggplot(data = plot_data_meta, aes(x=t, y=count, colour=state))+
  geom_line(alpha=0.8)+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 1, scales = "free_y")+
  labs(x="Time (Days)",
       y="Number of Individuals",
       colour="State")+
  theme_bw()
plot_meta

ggsave(filename = "meta_plot_7.pdf", 
       plot = plot_meta,
       device = "pdf",
       width = 7, 
       height = 8,
       path = "/Users/matthewhoyle/Github_R_projects/Plots/Hunter_Gatherer_models")

## Table showing extinction/transmission info for each patch

extinct_data_meta <- out_meta$data %>%
  as_tibble() %>%
  slice_max(t) %>%
  distinct() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N")),
         persist = case_when(state=="I" & count > 0 ~ T, 
                             state=="I" & count == 0 ~ F)) %>%
  drop_na() %>%
  select(patch, count, persist)
extinct_data_meta

beta_meta <- beta.ngm(beta_matrix)
paste0("Beta for whole system = ", beta_meta)


R0_meta <- R0ngm(nextgen_matrix)
paste0("R0 = ", R0_meta)


paste0("Actual number of infecteds at end of sim = ", sum(extinct_data_meta$count))
 # Total number of infecteds at the end of sim across all patches

sim_endpoint_meta <- as_tibble(out_meta$data) %>%
  slice_max(t) %>%
  distinct()


paste0("Did simulation run reach final endpoint?")
if (sim_endpoint_meta$t >= tf) {
  print("Yes")
} else {
  print("No")}

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta <- list()
sim_list_meta <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(lapply(
  seq_len(U), 
  function(x){ 
    c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
  }
))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))
  
  out_100_meta <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta <- out_100_meta$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta[[i]] <- sim_data_meta
}

sim_output_meta <- bind_rows(sim_list_meta)

# Summary table of endpoint data
sim_output_meta <- sim_output_meta %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta

# Make Summary Table of output
sim_summary_meta <- sim_output_meta %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,
            mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/100)
sim_summary_meta

Varying waining immunity

0 Days

#Collect parameters
parms_meta_0 <- parms_meta
parms_meta_0$omega <- 0


# Run simulations with the Direct method
set.seed(4)
out_meta_0 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_0,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_0 <- out_meta_0$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_0 <- ggplot(data = plot_data_meta_0, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_0

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_0 <- list()
sim_list_meta_0 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_0 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_0,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_0 <- out_100_meta_0$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_0[[i]] <- sim_data_meta_0
}

sim_output_meta_0 <- bind_rows(sim_list_meta_0)

# Summary table of endpoint data
sim_output_meta_0 <- sim_output_meta_0 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_0

# Make Summary Table of output
sim_summary_meta_0 <- sim_output_meta_0 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,
            mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 0)
sim_summary_meta_0

1 Day

#Collect parameters
parms_meta_1 <- parms_meta
parms_meta_1$omega <- 1


# Run simulations with the Direct method
set.seed(4)
out_meta_1 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_1,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_1 <- out_meta_1$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_1 <- ggplot(data = plot_data_meta_1, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_1

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_1 <- list()
sim_list_meta_1 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_1 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_1,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_1 <- out_100_meta_1$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_1[[i]] <- sim_data_meta_1
}

sim_output_meta_1 <- bind_rows(sim_list_meta_1)

# Summary table of endpoint data
sim_output_meta_1 <- sim_output_meta_1 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_1

# Make Summary Table of output
sim_summary_meta_1 <- sim_output_meta_1 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1)
sim_summary_meta_1

3 Days

#Collect parameters
parms_meta_3 <- parms_meta
parms_meta_3$omega <- 1/3


# Run simulations with the Direct method
set.seed(4)
out_meta_3 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_3,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_3 <- out_meta_3$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_3 <- ggplot(data = plot_data_meta_3, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_3

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_3 <- list()
sim_list_meta_3 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_3 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_3,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_3 <- out_100_meta_3$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_3[[i]] <- sim_data_meta_3
}

sim_output_meta_3 <- bind_rows(sim_list_meta_3)

# Summary table of endpoint data
sim_output_meta_3 <- sim_output_meta_3 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_3

# Make Summary Table of output
sim_summary_meta_3 <- sim_output_meta_3 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/3)
sim_summary_meta_3

7 Days

#Collect parameters
parms_meta_7 <- parms_meta
parms_meta_7$omega <- 1/7


# Run simulations with the Direct method
set.seed(4)
out_meta_7 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_7,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_7 <- out_meta_7$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_7 <- ggplot(data = plot_data_meta_7, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_7

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_7 <- list()
sim_list_meta_7 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_7 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_7,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_7 <- out_100_meta_7$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_7[[i]] <- sim_data_meta_7
}

sim_output_meta_7 <- bind_rows(sim_list_meta_7)

# Summary table of endpoint data
sim_output_meta_7 <- sim_output_meta_7 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_7

# Make Summary Table of output
sim_summary_meta_7 <- sim_output_meta_7 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/7)
sim_summary_meta_7

10 Days

#Collect parameters
parms_meta_10 <- parms_meta
parms_meta_10$omega <- 1/10

# Run simulations with the Direct method
set.seed(4)
out_meta_10 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_10,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_10 <- out_meta_10$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_10 <- ggplot(data = plot_data_meta_10, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_10

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_10 <- list()
sim_list_meta_10 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_10 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_10,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_10 <- out_100_meta_10$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_10[[i]] <- sim_data_meta_10
}

sim_output_meta_10 <- bind_rows(sim_list_meta_10)

# Summary table of endpoint data
sim_output_meta_10 <- sim_output_meta_10 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))

# Make Summary Table of output
sim_summary_meta_10 <- sim_output_meta_10 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/14)
sim_summary_meta_10

20 Days

#Collect parameters
parms_meta_20 <- parms_meta
parms_meta_20$omega <- 1/20


# Run simulations with the Direct method
set.seed(4)
out_meta_20 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_20,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_20 <- out_meta_20$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_20 <- ggplot(data = plot_data_meta_20, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_20

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_20 <- list()
sim_list_meta_20 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_20 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_20,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_20 <- out_100_meta_20$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_20[[i]] <- sim_data_meta_20
}

sim_output_meta_20 <- bind_rows(sim_list_meta_20)

# Summary table of endpoint data
sim_output_meta_20 <- sim_output_meta_20 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_20

# Make Summary Table of output
sim_summary_meta_20 <- sim_output_meta_20 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/20)
sim_summary_meta_20

30 Days

#Collect parameters
parms_meta_30 <- parms_meta
parms_meta_30$omega <- 1/30


# Run simulations with the Direct method
set.seed(4)
out_meta_30 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_30,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_30 <- out_meta_30$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_30 <- ggplot(data = plot_data_meta_30, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_30

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_30 <- list()
sim_list_meta_30 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_30 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_30,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_30 <- out_100_meta_30$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_30[[i]] <- sim_data_meta_30
}

sim_output_meta_30 <- bind_rows(sim_list_meta_30)

# Summary table of endpoint data
sim_output_meta_30 <- sim_output_meta_30 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_30

# Make Summary Table of output
sim_summary_meta_30 <- sim_output_meta_30 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,
            mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/30)
sim_summary_meta_30

40 Days

#Collect parameters
parms_meta_40 <- parms_meta
parms_meta_40$omega <- 1/40


# Run simulations with the Direct method
set.seed(4)
out_meta_40 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_40,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_40 <- out_meta_40$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_40 <- ggplot(data = plot_data_meta_40, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_40

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_40 <- list()
sim_list_meta_40 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_40 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_40,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_40 <- out_100_meta_40$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_40[[i]] <- sim_data_meta_40
}

sim_output_meta_40 <- bind_rows(sim_list_meta_40)

# Summary table of endpoint data
sim_output_meta_40 <- sim_output_meta_40 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_40

# Make Summary Table of output
sim_summary_meta_40 <- sim_output_meta_40 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/40)
sim_summary_meta_40

50 Days

#Collect parameters
parms_meta_50 <- parms_meta
parms_meta_50$omega <- 1/50


# Run simulations with the Direct method
set.seed(4)
out_meta_50 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_50,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_50 <- out_meta_50$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_50 <- ggplot(data = plot_data_meta_50, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_50

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_50 <- list()
sim_list_meta_50 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_50 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_50,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_50 <- out_100_meta_50$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_50[[i]] <- sim_data_meta_50
}

sim_output_meta_50 <- bind_rows(sim_list_meta_50)

# Summary table of endpoint data
sim_output_meta_50 <- sim_output_meta_50 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_50

# Make Summary Table of output
sim_summary_meta_50 <- sim_output_meta_50 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/50)
sim_summary_meta_50

60 Days

#Collect parameters
parms_meta_60 <- parms_meta
parms_meta_60$omega <- 1/60


# Run simulations with the Direct method
set.seed(4)
out_meta_60 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_60,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_60 <- out_meta_60$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_60 <- ggplot(data = plot_data_meta_60, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_60

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_60 <- list()
sim_list_meta_60 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_60 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_60,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_60 <- out_100_meta_60$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_60[[i]] <- sim_data_meta_60
}

sim_output_meta_60 <- bind_rows(sim_list_meta_60)

# Summary table of endpoint data
sim_output_meta_60 <- sim_output_meta_60 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_60

# Make Summary Table of output
sim_summary_meta_60 <- sim_output_meta_60 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/60)
sim_summary_meta_60

70 Days

#Collect parameters
parms_meta_70 <- parms_meta
parms_meta_70$omega <- 1/70


# Run simulations with the Direct method
set.seed(4)
out_meta_70 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_70,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_70 <- out_meta_70$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_70 <- ggplot(data = plot_data_meta_70, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_70

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_70 <- list()
sim_list_meta_70 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_70 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_70,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_70 <- out_100_meta_70$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_70[[i]] <- sim_data_meta_70
}

sim_output_meta_70 <- bind_rows(sim_list_meta_70)

# Summary table of endpoint data
sim_output_meta_70 <- sim_output_meta_70 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_70

# Make Summary Table of output
sim_summary_meta_70 <- sim_output_meta_70 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/70)
sim_summary_meta_70

80 Days

#Collect parameters
parms_meta_80 <- parms_meta
parms_meta_80$omega <- 1/80


# Run simulations with the Direct method
set.seed(4)
out_meta_80 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_80,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_80 <- out_meta_80$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_80 <- ggplot(data = plot_data_meta_80, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_80

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_80 <- list()
sim_list_meta_80 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_80 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_80,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_80 <- out_100_meta_80$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_80[[i]] <- sim_data_meta_80
}

sim_output_meta_80 <- bind_rows(sim_list_meta_80)

# Summary table of endpoint data
sim_output_meta_80 <- sim_output_meta_80 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_80

# Make Summary Table of output
sim_summary_meta_80 <- sim_output_meta_80 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/80)
sim_summary_meta_80

90 Days

#Collect parameters
parms_meta_90 <- parms_meta
parms_meta_90$omega <- 1/90


# Run simulations with the Direct method
set.seed(4)
out_meta_90 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_90,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_90 <- out_meta_90$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_90 <- ggplot(data = plot_data_meta_90, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_90

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_90 <- list()
sim_list_meta_90 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_90 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_90,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_90 <- out_100_meta_90$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_90[[i]] <- sim_data_meta_90
}

sim_output_meta_90 <- bind_rows(sim_list_meta_90)

# Summary table of endpoint data
sim_output_meta_90 <- sim_output_meta_90 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_90

# Make Summary Table of output
sim_summary_meta_90 <- sim_output_meta_90 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/90)
sim_summary_meta_90

180 Days

#Collect parameters
parms_meta_180 <- parms_meta
parms_meta_180$omega <- 1/180


# Run simulations with the Direct method
set.seed(20)
out_meta_180 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_180,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_180 <- out_meta_180$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_180 <- ggplot(data = plot_data_meta_180, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_180

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_180 <- list()
sim_list_meta_180 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_180 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_180,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_180 <- out_100_meta_180$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_180[[i]] <- sim_data_meta_180
}

sim_output_meta_180 <- bind_rows(sim_list_meta_180)

# Summary table of endpoint data
sim_output_meta_180 <- sim_output_meta_180 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_180

# Make Summary Table of output
sim_summary_meta_180 <- sim_output_meta_180 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/180)
sim_summary_meta_180

110 Days

#Collect parameters
parms_meta_110 <- parms_meta
parms_meta_110$omega <- 1/110


# Run simulations with the Direct method
set.seed(4)
out_meta_110 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_110,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_110 <- out_meta_110$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_110 <- ggplot(data = plot_data_meta_110, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_110

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_110 <- list()
sim_list_meta_110 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_110 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_110,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_110 <- out_100_meta_110$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_110[[i]] <- sim_data_meta_110
}

sim_output_meta_110 <- bind_rows(sim_list_meta_110)

# Summary table of endpoint data
sim_output_meta_110 <- sim_output_meta_110 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_110

# Make Summary Table of output
sim_summary_meta_110 <- sim_output_meta_110 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/110)
sim_summary_meta_110

120 Days

#Collect parameters
parms_meta_120 <- parms_meta
parms_meta_120$omega <- 1/120


# Run simulations with the Direct method
set.seed(4)
out_meta_120 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_120,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_120 <- out_meta_120$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_120 <- ggplot(data = plot_data_meta_120, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_120

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_120 <- list()
sim_list_meta_120 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_120 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_120,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_120 <- out_100_meta_120$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_120[[i]] <- sim_data_meta_120
}

sim_output_meta_120 <- bind_rows(sim_list_meta_120)

# Summary table of endpoint data
sim_output_meta_120 <- sim_output_meta_120 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_120

# Make Summary Table of output
sim_summary_meta_120 <- sim_output_meta_120 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/120)
sim_summary_meta_120

130 Days

#Collect parameters
parms_meta_130 <- parms_meta
parms_meta_130$omega <- 1/130


# Run simulations with the Direct method
set.seed(4)
out_meta_130 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_130,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_130 <- out_meta_130$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_130 <- ggplot(data = plot_data_meta_130, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_130

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_130 <- list()
sim_list_meta_130 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_130 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_130,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_130 <- out_100_meta_130$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_130[[i]] <- sim_data_meta_130
}

sim_output_meta_130 <- bind_rows(sim_list_meta_130)

# Summary table of endpoint data
sim_output_meta_130 <- sim_output_meta_130 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_130

# Make Summary Table of output
sim_summary_meta_130 <- sim_output_meta_130 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/130)
sim_summary_meta_130

150 Days

#Collect parameters
parms_meta_150 <- parms_meta
parms_meta_150$omega <- 1/150


# Run simulations with the Direct method
set.seed(4)
out_meta_150 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_150,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_150 <- out_meta_150$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_150 <- ggplot(data = plot_data_meta_150, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_150

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_150 <- list()
sim_list_meta_150 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_150 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_150,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_150 <- out_100_meta_150$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_150[[i]] <- sim_data_meta_150
}

sim_output_meta_150 <- bind_rows(sim_list_meta_150)

# Summary table of endpoint data
sim_output_meta_150 <- sim_output_meta_150 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_150

# Make Summary Table of output
sim_summary_meta_150 <- sim_output_meta_150 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/150)
sim_summary_meta_150

220 Days

#Collect parameters
parms_meta_220 <- parms_meta
parms_meta_220$omega <- 1/220


# Run simulations with the Direct method
set.seed(4)
out_meta_220 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_220,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_220 <- out_meta_220$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_220 <- ggplot(data = plot_data_meta_220, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_220

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_220 <- list()
sim_list_meta_220 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_220 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_220,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_220 <- out_100_meta_220$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_220[[i]] <- sim_data_meta_220
}

sim_output_meta_220 <- bind_rows(sim_list_meta_220)

# Summary table of endpoint data
sim_output_meta_220 <- sim_output_meta_220 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_220

# Make Summary Table of output
sim_summary_meta_220 <- sim_output_meta_220 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/220)
sim_summary_meta_220

270 Days

#Collect parameters
parms_meta_270 <- parms_meta
parms_meta_270$omega <- 1/270


# Run simulations with the Direct method
set.seed(4)
out_meta_270 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_270,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_270 <- out_meta_270$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_270 <- ggplot(data = plot_data_meta_270, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_270

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_270 <- list()
sim_list_meta_270 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_270 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_270,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_270 <- out_100_meta_270$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_270[[i]] <- sim_data_meta_270
}

sim_output_meta_270 <- bind_rows(sim_list_meta_270)

# Summary table of endpoint data
sim_output_meta_270 <- sim_output_meta_270 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_270

# Make Summary Table of output
sim_summary_meta_270 <- sim_output_meta_270 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/270)
sim_summary_meta_270

365 Days

#Collect parameters
parms_meta_365 <- parms_meta
parms_meta_365$omega <- 1/365


# Run simulations with the Direct method
set.seed(4)
out_meta_365 <- ssa(
  x0 = x0_meta,
  a = a_meta,
  nu = nu_meta,
  parms = parms_meta_365,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_meta_365 <- out_meta_365$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_meta_365 <- ggplot(data = plot_data_meta_365, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_meta_365

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_meta_365 <- list()
sim_list_meta_365 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize[x])
          }
        ))

names(x0_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_meta_365 <- ssa(
    x0 = x0_meta,
    a = a_meta,
    nu = nu_meta,
    parms = parms_meta_365,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_meta_365 <- out_100_meta_365$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_meta_365[[i]] <- sim_data_meta_365
}

sim_output_meta_365 <- bind_rows(sim_list_meta_365)

# Summary table of endpoint data
sim_output_meta_365 <- sim_output_meta_365 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_meta_365

# Make Summary Table of output
sim_summary_meta_365 <- sim_output_meta_365 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/365)
sim_summary_meta_365

Single

Results

waning_results_7 <- sim_summary_meta %>%
  bind_rows(sim_summary_meta_3) %>%
  bind_rows(sim_summary_meta_7) %>%
  bind_rows(sim_summary_meta_10) %>%
  bind_rows(sim_summary_meta_20) %>%
  bind_rows(sim_summary_meta_30) %>%
  bind_rows(sim_summary_meta_40) %>%
  bind_rows(sim_summary_meta_50) %>%
  bind_rows(sim_summary_meta_60) %>%
  bind_rows(sim_summary_meta_70) %>%
  bind_rows(sim_summary_meta_80) %>%
  bind_rows(sim_summary_meta_90) %>%
  bind_rows(sim_summary_meta_100) %>%
  bind_rows(sim_summary_meta_110) %>%
  bind_rows(sim_summary_meta_120) %>%
  bind_rows(sim_summary_meta_130) %>%
  bind_rows(sim_summary_meta_150) %>%
  bind_rows(sim_summary_meta_220) %>%
  bind_rows(sim_summary_meta_270) %>%
  bind_rows(sim_summary_meta_365) %>%
  mutate(immunity_duration = 1/omega) %>%
  arrange(immunity_duration) %>%
  mutate(model = "meta",
         patches = 7)

write_csv(waning_results_7, file = "/Users/matthewhoyle/Github_R_projects/Hunter_Gatherer_models/Results/waning_results_7.csv")

waning_results_7

ggplot(waning_results_7, aes(immunity_duration, sum_persist)) +
  geom_line()+
  geom_point()+
  theme_bw()

3-Patch Metapopulation Model

The same metapopulation SEIRS model was then used to model the dynamics of persistence in a 3-patch system and understand the effect of waning immunity.

###Set-up

# Define Paramenters
patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Patch size
U <- length(patchPopSize_3)                    # Number of patches
initial_infected <-  as.vector(rmultinom(1, 1, rep(0.5, U)))   # Initial infected (initial infected patch randomly generated)
initial_infected_patch <- which(initial_infected > 0)
simName <- "SIRS metapopulation model"       # Simulation name
tf <- 365*3                                   # Final time

# Agta Hunter-Gatherer contact rates
within_pop_contact = 1
between_pop_contact = 0.5/U     # normalised by number of patches 

#Create the named initial state vector for the U-patch system.

x0_3_meta <- unlist(lapply(
  seq_len(U), 
  function(i){ 
    c(patchPopSize_3[i] - initial_infected[i], initial_infected[i], 0, 0, patchPopSize_3[i])
  }
))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(i) paste0(c("S","E","I", "R", "N"), i)))


# Define the state change matrix for a single patch
nu_3_meta <- matrix(c( -1,  0,  0, +1, +1, -1,  0,  0,  0,  0, # S
                     +1, -1,  0,  0,  0,  0, -1,  0,  0,  0, # E
                      0, +1, -1,  0,  0,  0,  0, -1,  0, -1, # I
                      0,  0, +1, -1,  0,  0,  0,  0, -1,  0, # R 
                      0,  0,  0,  0, +1, -1 ,-1, -1, -1, -1), # N
             nrow=5,byrow=TRUE)

# Define propensity functions
# Mass-action
a_3_meta <-
  unlist(lapply(
    seq_len(U),
    function(patch) {
      i <- patch
      patches <- 1:U
      #j <- if (patch == 1) U else patch - 1
      other_patches <- patches[-i]
      patch_beta <- c()
      for(k in (1:(U-1))){
        patch_beta[k] = paste0("+(beta_", other_patches[k],i, "*I", other_patches[k], "/N", other_patches[k], ")*S", i)
      }
      c(
        paste0("(beta_", i, i, "*I", i,"/N", i, ")*S",i, paste0(patch_beta, collapse="")), # Infection
        paste0("sigma*E", i),                                       # Becomes infecious
        paste0("gamma*I", i),                                       # Recovery from infection
        paste0("omega*R", i),       # Loss of immunity
        paste0("mu*N", i),                             # Births
        paste0("mu*S", i),                                             # Deaths (S)
        paste0("mu*E", i),                                             # Deaths (E)
        paste0("mu*I", i),                                             # Deaths (I)
        paste0("mu*R", i),                                             # Deaths (R)
        paste0("alpha*I", i)                                           # Deaths from infection
        
      )
    }
  ))

Run Metapopulation Model

#Collect parameters
parms_3_meta <- list(
  sigma = 0.175,                          # E to I rate
  gamma = 0.2,                           # I to R rate
  omega = 1/100,                         # R to S rate
  mu = demo_sum$Birth_rate_daily_Mean,                            # Birth/death rate per person per day
  alpha = 1/1000) 

# Define transmission terms and populate next-generation matrix
beta <- 0.6

nextgen_3_matrix <- matrix(nrow = U, ncol = U, data = 0)
beta_3_matrix <- matrix(nrow = U, ncol = U, data = 0)


for(i in 1:U){
  for(j in 1:U){
    parms_3_meta[[paste0("beta_",i,i)]] = within_pop_contact*beta
    nextgen_3_matrix[i,i] = within_pop_contact*beta*(1/parms_3_meta$gamma)
    parms_3_meta[[paste0("beta_",j,i)]] = between_pop_contact*beta
    nextgen_3_matrix[j,i] = between_pop_contact*beta*(1/parms_3_meta$gamma)
    nextgen_3_matrix[i,j] = between_pop_contact*beta*(1/parms_3_meta$gamma)
    parms_3_meta[[paste0("beta_",j,j)]] = within_pop_contact*beta
    nextgen_3_matrix[j,j] = within_pop_contact*beta*(1/parms_3_meta$gamma)
    beta_3_matrix[i,i] = within_pop_contact*beta
    beta_3_matrix[j,i] = between_pop_contact*beta
    beta_3_matrix[i,j] = between_pop_contact*beta
    beta_3_matrix[j,j] = within_pop_contact*beta
  }
  parms_3_meta[[paste0("N", i)]] = patchPopSize_3[i]
}

# Run simulations with the Direct method
set.seed(25)
out_3_meta <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Plot
plot_data_3_meta <- out_3_meta$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta <- ggplot(data = plot_data_3_meta, aes(x=t, y=count, colour=state))+
  geom_line(alpha=0.8)+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 1, scales = "free_y")+
  labs(x="Time (Days)",
       y="Number of Individuals",
       colour="State")+
  theme_bw()
plot_3_meta

ggsave(filename = "meta_plot_3.pdf", 
       plot = plot_3_meta,
       device = "pdf",
       width = 7, 
       height = 8,
       path = "/Users/matthewhoyle/Github_R_projects/Plots/Hunter_Gatherer_models")

## Table showing extinction/transmission info for each patch

extinct_data_3_meta <- out_3_meta$data %>%
  as_tibble() %>%
  slice_max(t) %>%
  distinct() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N")),
         persist = case_when(state=="I" & count > 0 ~ T, 
                             state=="I" & count == 0 ~ F)) %>%
  drop_na() %>%
  select(patch, count, persist)
extinct_data_3_meta

beta_3_meta <- beta.ngm(beta_3_matrix)
paste0("Beta for whole system = ", beta_3_meta)


R0_3_meta <- R0ngm(nextgen_3_matrix)
paste0("R0 = ", R0_3_meta)


paste0("Actual number of infecteds at end of sim = ", sum(extinct_data_3_meta$count))
 # Total number of infecteds at the end of sim across all patches

sim_endpoint_3_meta <- as_tibble(out_3_meta$data) %>%
  slice_max(t) %>%
  distinct()


paste0("Did simulation run reach final endpoint?")
if (sim_endpoint_3_meta$t >= tf) {
  print("Yes")
} else {
  print("No")}

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta <- list()
sim_list_3_meta <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 7, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(lapply(
  seq_len(U), 
  function(x){ 
    c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
  }
))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))
  
  out_100_3_meta <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta <- out_100_3_meta$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta[[i]] <- sim_data_3_meta
}

sim_output_3_meta <- bind_rows(sim_list_3_meta)

# Summary table of endpoint data
sim_output_3_meta <- sim_output_3_meta %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta

# Make Summary Table of output
sim_summary_3_meta <- sim_output_3_meta %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,
            mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/100)
sim_summary_3_meta

Varying waining immunity

0 Days

#Collect parameters
parms_3_meta_0 <- parms_3_meta
parms_3_meta_0$omega <- 0


# Run simulations with the Direct method
set.seed(4)
out_3_meta_0 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_0,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_0 <- out_3_meta_0$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_0 <- ggplot(data = plot_data_3_meta_0, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_0

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_0 <- list()
sim_list_3_meta_0 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_0 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_0,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_0 <- out_100_3_meta_0$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_0[[i]] <- sim_data_3_meta_0
}

sim_output_3_meta_0 <- bind_rows(sim_list_3_meta_0)

# Summary table of endpoint data
sim_output_3_meta_0 <- sim_output_3_meta_0 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_0

# Make Summary Table of output
sim_summary_3_meta_0 <- sim_output_3_meta_0 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,
            mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 0)
sim_summary_3_meta_0

1 Day

#Collect parameters
parms_3_meta_1 <- parms_3_meta
parms_3_meta_1$omega <- 1


# Run simulations with the Direct method
set.seed(4)
out_3_meta_1 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_1,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_1 <- out_3_meta_1$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_1 <- ggplot(data = plot_data_3_meta_1, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_1

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_1 <- list()
sim_list_3_meta_1 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_1 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_1,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_1 <- out_100_3_meta_1$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_1[[i]] <- sim_data_3_meta_1
}

sim_output_3_meta_1 <- bind_rows(sim_list_3_meta_1)

# Summary table of endpoint data
sim_output_3_meta_1 <- sim_output_3_meta_1 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_1

# Make Summary Table of output
sim_summary_3_meta_1 <- sim_output_3_meta_1 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1)
sim_summary_3_meta_1

3 Days

#Collect parameters
parms_3_meta_3 <- parms_3_meta
parms_3_meta_3$omega <- 1/3


# Run simulations with the Direct method
set.seed(4)
out_3_meta_3 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_3,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_3 <- out_3_meta_3$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_3 <- ggplot(data = plot_data_3_meta_3, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_3

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_3 <- list()
sim_list_3_meta_3 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_3 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_3,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_3 <- out_100_3_meta_3$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_3[[i]] <- sim_data_3_meta_3
}

sim_output_3_meta_3 <- bind_rows(sim_list_3_meta_3)

# Summary table of endpoint data
sim_output_3_meta_3 <- sim_output_3_meta_3 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_3

# Make Summary Table of output
sim_summary_3_meta_3 <- sim_output_3_meta_3 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/3)
sim_summary_3_meta_3

7 Days

#Collect parameters
parms_3_meta_7 <- parms_3_meta
parms_3_meta_7$omega <- 1/7


# Run simulations with the Direct method
set.seed(4)
out_3_meta_7 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_7,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_7 <- out_3_meta_7$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_7 <- ggplot(data = plot_data_3_meta_7, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_7

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_7 <- list()
sim_list_3_meta_7 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_7 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_7,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_7 <- out_100_3_meta_7$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_7[[i]] <- sim_data_3_meta_7
}

sim_output_3_meta_7 <- bind_rows(sim_list_3_meta_7)

# Summary table of endpoint data
sim_output_3_meta_7 <- sim_output_3_meta_7 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_7

# Make Summary Table of output
sim_summary_3_meta_7 <- sim_output_3_meta_7 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/7)
sim_summary_3_meta_7

10 Days

#Collect parameters
parms_3_meta_10 <- parms_3_meta
parms_3_meta_10$omega <- 1/10

# Run simulations with the Direct method
set.seed(4)
out_3_meta_10 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_10,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_10 <- out_3_meta_10$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_10 <- ggplot(data = plot_data_3_meta_10, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_10

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_10 <- list()
sim_list_3_meta_10 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_10 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_10,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_10 <- out_100_3_meta_10$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_10[[i]] <- sim_data_3_meta_10
}

sim_output_3_meta_10 <- bind_rows(sim_list_3_meta_10)

# Summary table of endpoint data
sim_output_3_meta_10 <- sim_output_3_meta_10 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))

# Make Summary Table of output
sim_summary_3_meta_10 <- sim_output_3_meta_10 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/14)
sim_summary_3_meta_10

20 Days

#Collect parameters
parms_3_meta_20 <- parms_3_meta
parms_3_meta_20$omega <- 1/20


# Run simulations with the Direct method
set.seed(4)
out_3_meta_20 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_20,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_20 <- out_3_meta_20$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_20 <- ggplot(data = plot_data_3_meta_20, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_20

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_20 <- list()
sim_list_3_meta_20 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_20 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_20,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_20 <- out_100_3_meta_20$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_20[[i]] <- sim_data_3_meta_20
}

sim_output_3_meta_20 <- bind_rows(sim_list_3_meta_20)

# Summary table of endpoint data
sim_output_3_meta_20 <- sim_output_3_meta_20 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_20

# Make Summary Table of output
sim_summary_3_meta_20 <- sim_output_3_meta_20 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/20)
sim_summary_3_meta_20

30 Days

#Collect parameters
parms_3_meta_30 <- parms_3_meta
parms_3_meta_30$omega <- 1/30


# Run simulations with the Direct method
set.seed(4)
out_3_meta_30 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_30,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_30 <- out_3_meta_30$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_30 <- ggplot(data = plot_data_3_meta_30, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_30

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_30 <- list()
sim_list_3_meta_30 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_30 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_30,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_30 <- out_100_3_meta_30$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_30[[i]] <- sim_data_3_meta_30
}

sim_output_3_meta_30 <- bind_rows(sim_list_3_meta_30)

# Summary table of endpoint data
sim_output_3_meta_30 <- sim_output_3_meta_30 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_30

# Make Summary Table of output
sim_summary_3_meta_30 <- sim_output_3_meta_30 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,
            mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/30)
sim_summary_3_meta_30

40 Days

#Collect parameters
parms_3_meta_40 <- parms_3_meta
parms_3_meta_40$omega <- 1/40


# Run simulations with the Direct method
set.seed(4)
out_3_meta_40 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_40,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_40 <- out_3_meta_40$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_40 <- ggplot(data = plot_data_3_meta_40, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_40

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_40 <- list()
sim_list_3_meta_40 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_40 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_40,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_40 <- out_100_3_meta_40$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_40[[i]] <- sim_data_3_meta_40
}

sim_output_3_meta_40 <- bind_rows(sim_list_3_meta_40)

# Summary table of endpoint data
sim_output_3_meta_40 <- sim_output_3_meta_40 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_40

# Make Summary Table of output
sim_summary_3_meta_40 <- sim_output_3_meta_40 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/40)
sim_summary_3_meta_40

50 Days

#Collect parameters
parms_3_meta_50 <- parms_3_meta
parms_3_meta_50$omega <- 1/50


# Run simulations with the Direct method
set.seed(4)
out_3_meta_50 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_50,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_50 <- out_3_meta_50$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_50 <- ggplot(data = plot_data_3_meta_50, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_50

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_50 <- list()
sim_list_3_meta_50 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_50 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_50,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_50 <- out_100_3_meta_50$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_50[[i]] <- sim_data_3_meta_50
}

sim_output_3_meta_50 <- bind_rows(sim_list_3_meta_50)

# Summary table of endpoint data
sim_output_3_meta_50 <- sim_output_3_meta_50 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_50

# Make Summary Table of output
sim_summary_3_meta_50 <- sim_output_3_meta_50 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/50)
sim_summary_3_meta_50

60 Days

#Collect parameters
parms_3_meta_60 <- parms_3_meta
parms_3_meta_60$omega <- 1/60


# Run simulations with the Direct method
set.seed(4)
out_3_meta_60 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_60,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_60 <- out_3_meta_60$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_60 <- ggplot(data = plot_data_3_meta_60, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_60

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_60 <- list()
sim_list_3_meta_60 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_60 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_60,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_60 <- out_100_3_meta_60$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_60[[i]] <- sim_data_3_meta_60
}

sim_output_3_meta_60 <- bind_rows(sim_list_3_meta_60)

# Summary table of endpoint data
sim_output_3_meta_60 <- sim_output_3_meta_60 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_60

# Make Summary Table of output
sim_summary_3_meta_60 <- sim_output_3_meta_60 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/60)
sim_summary_3_meta_60

70 Days

#Collect parameters
parms_3_meta_70 <- parms_3_meta
parms_3_meta_70$omega <- 1/70


# Run simulations with the Direct method
set.seed(4)
out_3_meta_70 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_70,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_70 <- out_3_meta_70$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_70 <- ggplot(data = plot_data_3_meta_70, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_70

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_70 <- list()
sim_list_3_meta_70 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_70 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_70,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_70 <- out_100_3_meta_70$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_70[[i]] <- sim_data_3_meta_70
}

sim_output_3_meta_70 <- bind_rows(sim_list_3_meta_70)

# Summary table of endpoint data
sim_output_3_meta_70 <- sim_output_3_meta_70 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_70

# Make Summary Table of output
sim_summary_3_meta_70 <- sim_output_3_meta_70 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/70)
sim_summary_3_meta_70

80 Days

#Collect parameters
parms_3_meta_80 <- parms_3_meta
parms_3_meta_80$omega <- 1/80


# Run simulations with the Direct method
set.seed(4)
out_3_meta_80 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_80,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_80 <- out_3_meta_80$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_80 <- ggplot(data = plot_data_3_meta_80, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_80

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_80 <- list()
sim_list_3_meta_80 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_80 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_80,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_80 <- out_100_3_meta_80$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_80[[i]] <- sim_data_3_meta_80
}

sim_output_3_meta_80 <- bind_rows(sim_list_3_meta_80)

# Summary table of endpoint data
sim_output_3_meta_80 <- sim_output_3_meta_80 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_80

# Make Summary Table of output
sim_summary_3_meta_80 <- sim_output_3_meta_80 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/80)
sim_summary_3_meta_80

90 Days

#Collect parameters
parms_3_meta_90 <- parms_3_meta
parms_3_meta_90$omega <- 1/90


# Run simulations with the Direct method
set.seed(4)
out_3_meta_90 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_90,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_90 <- out_3_meta_90$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_90 <- ggplot(data = plot_data_3_meta_90, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_90

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_90 <- list()
sim_list_3_meta_90 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_90 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_90,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_90 <- out_100_3_meta_90$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_90[[i]] <- sim_data_3_meta_90
}

sim_output_3_meta_90 <- bind_rows(sim_list_3_meta_90)

# Summary table of endpoint data
sim_output_3_meta_90 <- sim_output_3_meta_90 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_90

# Make Summary Table of output
sim_summary_3_meta_90 <- sim_output_3_meta_90 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/90)
sim_summary_3_meta_90

180 Days

#Collect parameters
parms_3_meta_180 <- parms_3_meta
parms_3_meta_180$omega <- 1/180


# Run simulations with the Direct method
set.seed(20)
out_3_meta_180 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_180,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_180 <- out_3_meta_180$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_180 <- ggplot(data = plot_data_3_meta_180, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_180

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_180 <- list()
sim_list_3_meta_180 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_180 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_180,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_180 <- out_100_3_meta_180$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_180[[i]] <- sim_data_3_meta_180
}

sim_output_3_meta_180 <- bind_rows(sim_list_3_meta_180)

# Summary table of endpoint data
sim_output_3_meta_180 <- sim_output_3_meta_180 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_180

# Make Summary Table of output
sim_summary_3_meta_180 <- sim_output_3_meta_180 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/180)
sim_summary_3_meta_180

110 Days

#Collect parameters
parms_3_meta_110 <- parms_3_meta
parms_3_meta_110$omega <- 1/110


# Run simulations with the Direct method
set.seed(4)
out_3_meta_110 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_110,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_110 <- out_3_meta_110$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_110 <- ggplot(data = plot_data_3_meta_110, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_110

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_110 <- list()
sim_list_3_meta_110 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_110 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_110,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_110 <- out_100_3_meta_110$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_110[[i]] <- sim_data_3_meta_110
}

sim_output_3_meta_110 <- bind_rows(sim_list_3_meta_110)

# Summary table of endpoint data
sim_output_3_meta_110 <- sim_output_3_meta_110 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_110

# Make Summary Table of output
sim_summary_3_meta_110 <- sim_output_3_meta_110 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/110)
sim_summary_3_meta_110

120 Days

#Collect parameters
parms_3_meta_120 <- parms_3_meta
parms_3_meta_120$omega <- 1/120


# Run simulations with the Direct method
set.seed(4)
out_3_meta_120 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_120,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_120 <- out_3_meta_120$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_120 <- ggplot(data = plot_data_3_meta_120, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_120

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_120 <- list()
sim_list_3_meta_120 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_120 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_120,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_120 <- out_100_3_meta_120$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_120[[i]] <- sim_data_3_meta_120
}

sim_output_3_meta_120 <- bind_rows(sim_list_3_meta_120)

# Summary table of endpoint data
sim_output_3_meta_120 <- sim_output_3_meta_120 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_120

# Make Summary Table of output
sim_summary_3_meta_120 <- sim_output_3_meta_120 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/120)
sim_summary_3_meta_120

130 Days

#Collect parameters
parms_3_meta_130 <- parms_3_meta
parms_3_meta_130$omega <- 1/130


# Run simulations with the Direct method
set.seed(4)
out_3_meta_130 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_130,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_130 <- out_3_meta_130$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_130 <- ggplot(data = plot_data_3_meta_130, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_130

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_130 <- list()
sim_list_3_meta_130 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_130 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_130,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_130 <- out_100_3_meta_130$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_130[[i]] <- sim_data_3_meta_130
}

sim_output_3_meta_130 <- bind_rows(sim_list_3_meta_130)

# Summary table of endpoint data
sim_output_3_meta_130 <- sim_output_3_meta_130 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_130

# Make Summary Table of output
sim_summary_3_meta_130 <- sim_output_3_meta_130 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/130)
sim_summary_3_meta_130

150 Days

#Collect parameters
parms_3_meta_150 <- parms_3_meta
parms_3_meta_150$omega <- 1/150


# Run simulations with the Direct method
set.seed(4)
out_3_meta_150 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_150,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_150 <- out_3_meta_150$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_150 <- ggplot(data = plot_data_3_meta_150, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_150

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_150 <- list()
sim_list_3_meta_150 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_150 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_150,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_150 <- out_100_3_meta_150$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_150[[i]] <- sim_data_3_meta_150
}

sim_output_3_meta_150 <- bind_rows(sim_list_3_meta_150)

# Summary table of endpoint data
sim_output_3_meta_150 <- sim_output_3_meta_150 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_150

# Make Summary Table of output
sim_summary_3_meta_150 <- sim_output_3_meta_150 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/150)
sim_summary_3_meta_150

220 Days

#Collect parameters
parms_3_meta_220 <- parms_3_meta
parms_3_meta_220$omega <- 1/220


# Run simulations with the Direct method
set.seed(4)
out_3_meta_220 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_220,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_220 <- out_3_meta_220$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_220 <- ggplot(data = plot_data_3_meta_220, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_220

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_220 <- list()
sim_list_3_meta_220 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_220 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_220,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_220 <- out_100_3_meta_220$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_220[[i]] <- sim_data_3_meta_220
}

sim_output_3_meta_220 <- bind_rows(sim_list_3_meta_220)

# Summary table of endpoint data
sim_output_3_meta_220 <- sim_output_3_meta_220 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_220

# Make Summary Table of output
sim_summary_3_meta_220 <- sim_output_3_meta_220 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/220)
sim_summary_3_meta_220

270 Days

#Collect parameters
parms_3_meta_270 <- parms_3_meta
parms_3_meta_270$omega <- 1/270


# Run simulations with the Direct method
set.seed(4)
out_3_meta_270 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_270,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_270 <- out_3_meta_270$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_270 <- ggplot(data = plot_data_3_meta_270, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_270

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_270 <- list()
sim_list_3_meta_270 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_270 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_270,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_270 <- out_100_3_meta_270$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_270[[i]] <- sim_data_3_meta_270
}

sim_output_3_meta_270 <- bind_rows(sim_list_3_meta_270)

# Summary table of endpoint data
sim_output_3_meta_270 <- sim_output_3_meta_270 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_270

# Make Summary Table of output
sim_summary_3_meta_270 <- sim_output_3_meta_270 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/270)
sim_summary_3_meta_270

365 Days

#Collect parameters
parms_3_meta_365 <- parms_3_meta
parms_3_meta_365$omega <- 1/365


# Run simulations with the Direct method
set.seed(4)
out_3_meta_365 <- ssa(
  x0 = x0_3_meta,
  a = a_3_meta,
  nu = nu_3_meta,
  parms = parms_3_meta_365,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_3_meta_365 <- out_3_meta_365$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_3_meta_365 <- ggplot(data = plot_data_3_meta_365, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_3_meta_365

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_3_meta_365 <- list()
sim_list_3_meta_365 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 3, replace = TRUE)    # Sample different patch sizes for each sim
  x0_3_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_3_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_3_meta_365 <- ssa(
    x0 = x0_3_meta,
    a = a_3_meta,
    nu = nu_3_meta,
    parms = parms_3_meta_365,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_3_meta_365 <- out_100_3_meta_365$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_3_meta_365[[i]] <- sim_data_3_meta_365
}

sim_output_3_meta_365 <- bind_rows(sim_list_3_meta_365)

# Summary table of endpoint data
sim_output_3_meta_365 <- sim_output_3_meta_365 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_3_meta_365

# Make Summary Table of output
sim_summary_3_meta_365 <- sim_output_3_meta_365 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/365)
sim_summary_3_meta_365

Single

Results

waning_results_3 <- sim_summary_3_meta %>%
  bind_rows(sim_summary_3_meta_3) %>%
  bind_rows(sim_summary_3_meta_7) %>%
  bind_rows(sim_summary_3_meta_10) %>%
  bind_rows(sim_summary_3_meta_20) %>%
  bind_rows(sim_summary_3_meta_30) %>%
  bind_rows(sim_summary_3_meta_40) %>%
  bind_rows(sim_summary_3_meta_50) %>%
  bind_rows(sim_summary_3_meta_60) %>%
  bind_rows(sim_summary_3_meta_70) %>%
  bind_rows(sim_summary_3_meta_80) %>%
  bind_rows(sim_summary_3_meta_90) %>%
  bind_rows(sim_summary_3_meta_180) %>%
  bind_rows(sim_summary_3_meta_110) %>%
  bind_rows(sim_summary_3_meta_120) %>%
  bind_rows(sim_summary_3_meta_130) %>%
  bind_rows(sim_summary_3_meta_150) %>%
  bind_rows(sim_summary_3_meta_220) %>%
  bind_rows(sim_summary_3_meta_270) %>%
  bind_rows(sim_summary_3_meta_365) %>%
  mutate(immunity_duration = 1/omega) %>%
  arrange(immunity_duration) %>%
  mutate(model = "meta",
         patches = 3)

write_csv(waning_results_3, file = "/Users/matthewhoyle/Github_R_projects/Hunter_Gatherer_models/Results/waning_results_3.csv")

waning_results_3

ggplot(waning_results_3, aes(immunity_duration, sum_persist)) +
  geom_line()+
  geom_point()+
  theme_bw()

14 Metapopulation Model

The same metapopulation SEIRS model was then used to model the dynamics of persistence in a 14-patch system and understand the effect of waning immunity.

###Set-up

# Define Paramenters
patchPopSize_14 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Patch size
U <- length(patchPopSize_14)                    # Number of patches
initial_infected <-  as.vector(rmultinom(1, 1, rep(0.5, U)))   # Initial infected (initial infected patch randomly generated)
initial_infected_patch <- which(initial_infected > 0)
simName <- "SIRS metapopulation model"       # Simulation name
tf <- 365*3                                   # Final time

# Agta Hunter-Gatherer contact rates
within_pop_contact = 1
between_pop_contact = 0.5/U     # normalised by number of patches 

#Create the named initial state vector for the U-patch system.

x0_14_meta <- unlist(lapply(
  seq_len(U), 
  function(i){ 
    c(patchPopSize_14[i] - initial_infected[i], initial_infected[i], 0, 0, patchPopSize_14[i])
  }
))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(i) paste0(c("S","E","I", "R", "N"), i)))


# Define the state change matrix for a single patch
nu_14_meta <- matrix(c( -1,  0,  0, +1, +1, -1,  0,  0,  0,  0, # S
                     +1, -1,  0,  0,  0,  0, -1,  0,  0,  0, # E
                      0, +1, -1,  0,  0,  0,  0, -1,  0, -1, # I
                      0,  0, +1, -1,  0,  0,  0,  0, -1,  0, # R 
                      0,  0,  0,  0, +1, -1 ,-1, -1, -1, -1), # N
             nrow=5,byrow=TRUE)

# Define propensity functions
# Mass-action
a_14_meta <-
  unlist(lapply(
    seq_len(U),
    function(patch) {
      i <- patch
      patches <- 1:U
      #j <- if (patch == 1) U else patch - 1
      other_patches <- patches[-i]
      patch_beta <- c()
      for(k in (1:(U-1))){
        patch_beta[k] = paste0("+(beta_", other_patches[k],i, "*I", other_patches[k], "/N", other_patches[k], ")*S", i)
      }
      c(
        paste0("(beta_", i, i, "*I", i,"/N", i, ")*S",i, paste0(patch_beta, collapse="")), # Infection
        paste0("sigma*E", i),                                       # Becomes infecious
        paste0("gamma*I", i),                                       # Recovery from infection
        paste0("omega*R", i),       # Loss of immunity
        paste0("mu*N", i),                             # Births
        paste0("mu*S", i),                                             # Deaths (S)
        paste0("mu*E", i),                                             # Deaths (E)
        paste0("mu*I", i),                                             # Deaths (I)
        paste0("mu*R", i),                                             # Deaths (R)
        paste0("alpha*I", i)                                           # Deaths from infection
        
      )
    }
  ))

Run Metapopulation Model

#Collect parameters
parms_14_meta <- list(
  sigma = 0.175,                          # E to I rate
  gamma = 0.2,                           # I to R rate
  omega = 1/100,                         # R to S rate
  mu = demo_sum$Birth_rate_daily_Mean,                            # Birth/death rate per person per day
  alpha = 1/1000) 

# Define transmission terms and populate next-generation matrix
beta <- 0.6

nextgen_14_matrix <- matrix(nrow = U, ncol = U, data = 0)
beta_14_matrix <- matrix(nrow = U, ncol = U, data = 0)


for(i in 1:U){
  for(j in 1:U){
    parms_14_meta[[paste0("beta_",i,i)]] = within_pop_contact*beta
    nextgen_14_matrix[i,i] = within_pop_contact*beta*(1/parms_14_meta$gamma)
    parms_14_meta[[paste0("beta_",j,i)]] = between_pop_contact*beta
    nextgen_14_matrix[j,i] = between_pop_contact*beta*(1/parms_14_meta$gamma)
    nextgen_14_matrix[i,j] = between_pop_contact*beta*(1/parms_14_meta$gamma)
    parms_14_meta[[paste0("beta_",j,j)]] = within_pop_contact*beta
    nextgen_14_matrix[j,j] = within_pop_contact*beta*(1/parms_14_meta$gamma)
    beta_14_matrix[i,i] = within_pop_contact*beta
    beta_14_matrix[j,i] = between_pop_contact*beta
    beta_14_matrix[i,j] = between_pop_contact*beta
    beta_14_matrix[j,j] = within_pop_contact*beta
  }
  parms_14_meta[[paste0("N", i)]] = patchPopSize_14[i]
}

# Run simulations with the Direct method
set.seed(25)
out_14_meta <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Plot
plot_data_14_meta <- out_14_meta$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta <- ggplot(data = plot_data_14_meta, aes(x=t, y=count, colour=state))+
  geom_line(alpha=0.8)+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time (Days)",
       y="Number of Individuals",
       colour="State")+
  theme_bw()
plot_14_meta

ggsave(filename = "meta_14_plot.pdf", 
       plot = plot_14_meta,
       device = "pdf",
       width = 7, 
       height = 8,
       path = "/Users/matthewhoyle/Github_R_projects/Plots/Hunter_Gatherer_models")

## Table showing extinction/transmission info for each patch

extinct_data_14_meta <- out_14_meta$data %>%
  as_tibble() %>%
  slice_max(t) %>%
  distinct() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N")),
         persist = case_when(state=="I" & count > 0 ~ T, 
                             state=="I" & count == 0 ~ F)) %>%
  drop_na() %>%
  select(patch, count, persist)
extinct_data_14_meta

beta_14_meta <- beta.ngm(beta_14_matrix)
paste0("Beta for whole system = ", beta_14_meta)


R0_14_meta <- R0ngm(nextgen_14_matrix)
paste0("R0 = ", R0_14_meta)


paste0("Actual number of infecteds at end of sim = ", sum(extinct_data_14_meta$count))
 # Total number of infecteds at the end of sim across all patches

sim_endpoint_14_meta <- as_tibble(out_14_meta$data) %>%
  slice_max(t) %>%
  distinct()


paste0("Did simulation run reach final endpoint?")
if (sim_endpoint_14_meta$t >= tf) {
  print("Yes")
} else {
  print("No")}

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta <- list()
sim_list_14_meta <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_14 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(lapply(
  seq_len(U), 
  function(x){ 
    c(patchPopSize_14[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_14[x])
  }
))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))
  
  out_100_14_meta <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta <- out_100_14_meta$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta[[i]] <- sim_data_14_meta
}

sim_output_14_meta <- bind_rows(sim_list_14_meta)

# Summary table of endpoint data
sim_output_14_meta <- sim_output_14_meta %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta

# Make Summary Table of output
sim_summary_14_meta <- sim_output_14_meta %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,
            mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/100)
sim_summary_14_meta

Varying waining immunity

0 Days

#Collect parameters
parms_14_meta_0 <- parms_14_meta
parms_14_meta_0$omega <- 0


# Run simulations with the Direct method
set.seed(4)
out_14_meta_0 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_0,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_0 <- out_14_meta_0$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_0 <- ggplot(data = plot_data_14_meta_0, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_0

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_0 <- list()
sim_list_14_meta_0 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_0 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_0,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_0 <- out_100_14_meta_0$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_0[[i]] <- sim_data_14_meta_0
}

sim_output_14_meta_0 <- bind_rows(sim_list_14_meta_0)

# Summary table of endpoint data
sim_output_14_meta_0 <- sim_output_14_meta_0 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_0

# Make Summary Table of output
sim_summary_14_meta_0 <- sim_output_14_meta_0 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,
            mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 0)
sim_summary_14_meta_0

10 Days

#Collect parameters
parms_14_meta_10 <- parms_14_meta
parms_14_meta_10$omega <- 1/10

# Run simulations with the Direct method
set.seed(4)
out_14_meta_10 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_10,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_10 <- out_14_meta_10$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_10 <- ggplot(data = plot_data_14_meta_10, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_10

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_10 <- list()
sim_list_14_meta_10 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_10 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_10,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_10 <- out_100_14_meta_10$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_10[[i]] <- sim_data_14_meta_10
}

sim_output_14_meta_10 <- bind_rows(sim_list_14_meta_10)

# Summary table of endpoint data
sim_output_14_meta_10 <- sim_output_14_meta_10 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))

# Make Summary Table of output
sim_summary_14_meta_10 <- sim_output_14_meta_10 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/10)
sim_summary_14_meta_10

20 Days

#Collect parameters
parms_14_meta_20 <- parms_14_meta
parms_14_meta_20$omega <- 1/20


# Run simulations with the Direct method
set.seed(4)
out_14_meta_20 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_20,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_20 <- out_14_meta_20$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_20 <- ggplot(data = plot_data_14_meta_20, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_20

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_20 <- list()
sim_list_14_meta_20 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_20 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_20,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_20 <- out_100_14_meta_20$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_20[[i]] <- sim_data_14_meta_20
}

sim_output_14_meta_20 <- bind_rows(sim_list_14_meta_20)

# Summary table of endpoint data
sim_output_14_meta_20 <- sim_output_14_meta_20 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_20

# Make Summary Table of output
sim_summary_14_meta_20 <- sim_output_14_meta_20 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/20)
sim_summary_14_meta_20

30 Days

#Collect parameters
parms_14_meta_30 <- parms_14_meta
parms_14_meta_30$omega <- 1/30


# Run simulations with the Direct method
set.seed(4)
out_14_meta_30 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_30,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_30 <- out_14_meta_30$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_30 <- ggplot(data = plot_data_14_meta_30, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_30

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_30 <- list()
sim_list_14_meta_30 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_30 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_30,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_30 <- out_100_14_meta_30$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_30[[i]] <- sim_data_14_meta_30
}

sim_output_14_meta_30 <- bind_rows(sim_list_14_meta_30)

# Summary table of endpoint data
sim_output_14_meta_30 <- sim_output_14_meta_30 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_30

# Make Summary Table of output
sim_summary_14_meta_30 <- sim_output_14_meta_30 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,
            mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/30)
sim_summary_14_meta_30

40 Days

#Collect parameters
parms_14_meta_40 <- parms_14_meta
parms_14_meta_40$omega <- 1/40


# Run simulations with the Direct method
set.seed(4)
out_14_meta_40 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_40,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_40 <- out_14_meta_40$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_40 <- ggplot(data = plot_data_14_meta_40, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_40

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_40 <- list()
sim_list_14_meta_40 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_40 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_40,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_40 <- out_100_14_meta_40$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_40[[i]] <- sim_data_14_meta_40
}

sim_output_14_meta_40 <- bind_rows(sim_list_14_meta_40)

# Summary table of endpoint data
sim_output_14_meta_40 <- sim_output_14_meta_40 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_40

# Make Summary Table of output
sim_summary_14_meta_40 <- sim_output_14_meta_40 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/40)
sim_summary_14_meta_40

50 Days

#Collect parameters
parms_14_meta_50 <- parms_14_meta
parms_14_meta_50$omega <- 1/50


# Run simulations with the Direct method
set.seed(4)
out_14_meta_50 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_50,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_50 <- out_14_meta_50$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_50 <- ggplot(data = plot_data_14_meta_50, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_50

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_50 <- list()
sim_list_14_meta_50 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_50 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_50,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_50 <- out_100_14_meta_50$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_50[[i]] <- sim_data_14_meta_50
}

sim_output_14_meta_50 <- bind_rows(sim_list_14_meta_50)

# Summary table of endpoint data
sim_output_14_meta_50 <- sim_output_14_meta_50 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_50

# Make Summary Table of output
sim_summary_14_meta_50 <- sim_output_14_meta_50 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/50)
sim_summary_14_meta_50

60 Days

#Collect parameters
parms_14_meta_60 <- parms_14_meta
parms_14_meta_60$omega <- 1/60


# Run simulations with the Direct method
set.seed(4)
out_14_meta_60 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_60,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_60 <- out_14_meta_60$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_60 <- ggplot(data = plot_data_14_meta_60, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_60

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_60 <- list()
sim_list_14_meta_60 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_60 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_60,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_60 <- out_100_14_meta_60$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_60[[i]] <- sim_data_14_meta_60
}

sim_output_14_meta_60 <- bind_rows(sim_list_14_meta_60)

# Summary table of endpoint data
sim_output_14_meta_60 <- sim_output_14_meta_60 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_60

# Make Summary Table of output
sim_summary_14_meta_60 <- sim_output_14_meta_60 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/60)
sim_summary_14_meta_60

70 Days

#Collect parameters
parms_14_meta_70 <- parms_14_meta
parms_14_meta_70$omega <- 1/70


# Run simulations with the Direct method
set.seed(4)
out_14_meta_70 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_70,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_70 <- out_14_meta_70$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_70 <- ggplot(data = plot_data_14_meta_70, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_70

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_70 <- list()
sim_list_14_meta_70 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_70 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_70,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_70 <- out_100_14_meta_70$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_70[[i]] <- sim_data_14_meta_70
}

sim_output_14_meta_70 <- bind_rows(sim_list_14_meta_70)

# Summary table of endpoint data
sim_output_14_meta_70 <- sim_output_14_meta_70 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_70

# Make Summary Table of output
sim_summary_14_meta_70 <- sim_output_14_meta_70 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/70)
sim_summary_14_meta_70

80 Days

#Collect parameters
parms_14_meta_80 <- parms_14_meta
parms_14_meta_80$omega <- 1/80


# Run simulations with the Direct method
set.seed(4)
out_14_meta_80 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_80,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_80 <- out_14_meta_80$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_80 <- ggplot(data = plot_data_14_meta_80, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_80

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_80 <- list()
sim_list_14_meta_80 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_80 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_80,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_80 <- out_100_14_meta_80$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_80[[i]] <- sim_data_14_meta_80
}

sim_output_14_meta_80 <- bind_rows(sim_list_14_meta_80)

# Summary table of endpoint data
sim_output_14_meta_80 <- sim_output_14_meta_80 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_80

# Make Summary Table of output
sim_summary_14_meta_80 <- sim_output_14_meta_80 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/80)
sim_summary_14_meta_80

90 Days

#Collect parameters
parms_14_meta_90 <- parms_14_meta
parms_14_meta_90$omega <- 1/90


# Run simulations with the Direct method
set.seed(4)
out_14_meta_90 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_90,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_90 <- out_14_meta_90$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_90 <- ggplot(data = plot_data_14_meta_90, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_90

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_90 <- list()
sim_list_14_meta_90 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_90 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_90,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_90 <- out_100_14_meta_90$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_90[[i]] <- sim_data_14_meta_90
}

sim_output_14_meta_90 <- bind_rows(sim_list_14_meta_90)

# Summary table of endpoint data
sim_output_14_meta_90 <- sim_output_14_meta_90 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_90

# Make Summary Table of output
sim_summary_14_meta_90 <- sim_output_14_meta_90 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/90)
sim_summary_14_meta_90

180 Days

#Collect parameters
parms_14_meta_180 <- parms_14_meta
parms_14_meta_180$omega <- 1/180


# Run simulations with the Direct method
set.seed(20)
out_14_meta_180 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_180,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_180 <- out_14_meta_180$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_180 <- ggplot(data = plot_data_14_meta_180, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_180

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_180 <- list()
sim_list_14_meta_180 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_180 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_180,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_180 <- out_100_14_meta_180$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_180[[i]] <- sim_data_14_meta_180
}

sim_output_14_meta_180 <- bind_rows(sim_list_14_meta_180)

# Summary table of endpoint data
sim_output_14_meta_180 <- sim_output_14_meta_180 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_180

# Make Summary Table of output
sim_summary_14_meta_180 <- sim_output_14_meta_180 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/180)
sim_summary_14_meta_180

110 Days

#Collect parameters
parms_14_meta_110 <- parms_14_meta
parms_14_meta_110$omega <- 1/110


# Run simulations with the Direct method
set.seed(4)
out_14_meta_110 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_110,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_110 <- out_14_meta_110$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_110 <- ggplot(data = plot_data_14_meta_110, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_110

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_110 <- list()
sim_list_14_meta_110 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_110 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_110,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_110 <- out_100_14_meta_110$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_110[[i]] <- sim_data_14_meta_110
}

sim_output_14_meta_110 <- bind_rows(sim_list_14_meta_110)

# Summary table of endpoint data
sim_output_14_meta_110 <- sim_output_14_meta_110 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_110

# Make Summary Table of output
sim_summary_14_meta_110 <- sim_output_14_meta_110 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/110)
sim_summary_14_meta_110

120 Days

#Collect parameters
parms_14_meta_120 <- parms_14_meta
parms_14_meta_120$omega <- 1/120


# Run simulations with the Direct method
set.seed(4)
out_14_meta_120 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_120,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_120 <- out_14_meta_120$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_120 <- ggplot(data = plot_data_14_meta_120, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_120

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_120 <- list()
sim_list_14_meta_120 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_120 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_120,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_120 <- out_100_14_meta_120$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_120[[i]] <- sim_data_14_meta_120
}

sim_output_14_meta_120 <- bind_rows(sim_list_14_meta_120)

# Summary table of endpoint data
sim_output_14_meta_120 <- sim_output_14_meta_120 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_120

# Make Summary Table of output
sim_summary_14_meta_120 <- sim_output_14_meta_120 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/120)
sim_summary_14_meta_120

130 Days

#Collect parameters
parms_14_meta_130 <- parms_14_meta
parms_14_meta_130$omega <- 1/130


# Run simulations with the Direct method
set.seed(4)
out_14_meta_130 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_130,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_130 <- out_14_meta_130$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_130 <- ggplot(data = plot_data_14_meta_130, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_130

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_130 <- list()
sim_list_14_meta_130 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_130 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_130,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_130 <- out_100_14_meta_130$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_130[[i]] <- sim_data_14_meta_130
}

sim_output_14_meta_130 <- bind_rows(sim_list_14_meta_130)

# Summary table of endpoint data
sim_output_14_meta_130 <- sim_output_14_meta_130 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_130

# Make Summary Table of output
sim_summary_14_meta_130 <- sim_output_14_meta_130 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/130)
sim_summary_14_meta_130

150 Days

#Collect parameters
parms_14_meta_150 <- parms_14_meta
parms_14_meta_150$omega <- 1/150


# Run simulations with the Direct method
set.seed(4)
out_14_meta_150 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_150,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_150 <- out_14_meta_150$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_150 <- ggplot(data = plot_data_14_meta_150, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_150

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_150 <- list()
sim_list_14_meta_150 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_150 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_150,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_150 <- out_100_14_meta_150$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_150[[i]] <- sim_data_14_meta_150
}

sim_output_14_meta_150 <- bind_rows(sim_list_14_meta_150)

# Summary table of endpoint data
sim_output_14_meta_150 <- sim_output_14_meta_150 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_150

# Make Summary Table of output
sim_summary_14_meta_150 <- sim_output_14_meta_150 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/150)
sim_summary_14_meta_150

220 Days

#Collect parameters
parms_14_meta_220 <- parms_14_meta
parms_14_meta_220$omega <- 1/220


# Run simulations with the Direct method
set.seed(4)
out_14_meta_220 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_220,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_220 <- out_14_meta_220$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_220 <- ggplot(data = plot_data_14_meta_220, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_220

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_220 <- list()
sim_list_14_meta_220 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_220 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_220,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_220 <- out_100_14_meta_220$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_220[[i]] <- sim_data_14_meta_220
}

sim_output_14_meta_220 <- bind_rows(sim_list_14_meta_220)

# Summary table of endpoint data
sim_output_14_meta_220 <- sim_output_14_meta_220 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_220

# Make Summary Table of output
sim_summary_14_meta_220 <- sim_output_14_meta_220 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/220)
sim_summary_14_meta_220

270 Days

#Collect parameters
parms_14_meta_270 <- parms_14_meta
parms_14_meta_270$omega <- 1/270


# Run simulations with the Direct method
set.seed(4)
out_14_meta_270 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_270,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_270 <- out_14_meta_270$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_270 <- ggplot(data = plot_data_14_meta_270, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_270

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_270 <- list()
sim_list_14_meta_270 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_270 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_270,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_270 <- out_100_14_meta_270$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_270[[i]] <- sim_data_14_meta_270
}

sim_output_14_meta_270 <- bind_rows(sim_list_14_meta_270)

# Summary table of endpoint data
sim_output_14_meta_270 <- sim_output_14_meta_270 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_270

# Make Summary Table of output
sim_summary_14_meta_270 <- sim_output_14_meta_270 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/270)
sim_summary_14_meta_270

365 Days

#Collect parameters
parms_14_meta_365 <- parms_14_meta
parms_14_meta_365$omega <- 1/365


# Run simulations with the Direct method
set.seed(4)
out_14_meta_365 <- ssa(
  x0 = x0_14_meta,
  a = a_14_meta,
  nu = nu_14_meta,
  parms = parms_14_meta_365,
  tf = tf,
  method = ssa.d(),
  simName = simName,
  verbose = FALSE,
  consoleInterval = 1
)


## Extra Plots
plot_data_14_meta_365 <- out_14_meta_365$data %>%
  as_tibble() %>%
  pivot_longer(!t, names_to = "ID", values_to = "count") %>%
  separate(ID, 
           into = c("state", "patch"), 
           sep = "(?<=[A-Za-z])(?=[0-9])") %>%
  mutate(state = factor(state, levels = c("S", "E", "I", "R", "N"))) #%>%
  #filter(state != "N")

plot_14_meta_365 <- ggplot(data = plot_data_14_meta_365, aes(x=t, y=count, colour=state))+
  geom_line()+
  facet_wrap(~factor(patch, levels = unique(patch)) ,ncol = 3, scales = "free_y")+
  labs(x="Time",
       y="Frequency")+
  theme_bw()
plot_14_meta_365

## Run multiple simulations and saving output
num_sims <- 1000
sim_list_14_meta_365 <- list()
sim_list_14_meta_365 <- vector("list", length = num_sims)

for (i in 1:num_sims){
  set.seed(i)
  
  patchPopSize_3 <-     sample(camps.data$camp_total, 14, replace = TRUE)    # Sample different patch sizes for each sim
  x0_14_meta <- unlist(
      lapply(
        seq_len(U),
        function(x){
          c(patchPopSize_3[x] - initial_infected[x], initial_infected[x], 0, 0, patchPopSize_3[x])
          }
        ))

names(x0_14_meta) <- unlist(lapply(seq_len(U), function(x) paste0(c("S","E","I", "R", "N"), x)))

  out_100_14_meta_365 <- ssa(
    x0 = x0_14_meta,
    a = a_14_meta,
    nu = nu_14_meta,
    parms = parms_14_meta_365,
    tf = tf,
    method = ssa.d(),
    simName = simName,
    verbose = FALSE,
    consoleInterval = 1
  )

  
# Extract Final time point from output data
  sim_data_14_meta_365 <- out_100_14_meta_365$data %>%
    as_tibble() %>%
    slice_max(t) %>%
    distinct() %>%
    pivot_longer(!t, names_to = "ID", values_to = "count") %>%
    separate(ID, 
             into = c("state", "patch"), 
             sep = "(?<=[A-Za-z])(?=[0-9])") %>%
    pivot_wider(names_from = state, values_from = count) %>%
    mutate(persist = case_when(I > 0 ~ T, 
                               I == 0 ~ F),
           sim = i) %>%
    select(sim, patch, I, N, persist)
  
  sim_list_14_meta_365[[i]] <- sim_data_14_meta_365
}

sim_output_14_meta_365 <- bind_rows(sim_list_14_meta_365)

# Summary table of endpoint data
sim_output_14_meta_365 <- sim_output_14_meta_365 %>%
  group_by(sim) %>%
  summarise(total_I = sum(I), 
            total_N = sum(N)) %>%
  mutate(percent_persist = total_I/(total_N)*100,
         persist = case_when(total_I > 0 ~ T, 
                               total_I == 0 ~ F))
sim_output_14_meta_365

# Make Summary Table of output
sim_summary_14_meta_365 <- sim_output_14_meta_365 %>%
  summarise(mean_infecteds = mean(total_I),
            sum_persist = (sum(persist, na.rm = T)/num_sims)*100,              mean_percent_infected = mean(percent_persist)) %>%
  mutate(omega = 1/365)
sim_summary_14_meta_365

Results

waning_results_14 <- sim_summary_14_meta %>%
  bind_rows(sim_summary_14_meta_10) %>%
  bind_rows(sim_summary_14_meta_20) %>%
  bind_rows(sim_summary_14_meta_30) %>%
  bind_rows(sim_summary_14_meta_40) %>%
  bind_rows(sim_summary_14_meta_50) %>%
  bind_rows(sim_summary_14_meta_60) %>%
  bind_rows(sim_summary_14_meta_70) %>%
  bind_rows(sim_summary_14_meta_80) %>%
  bind_rows(sim_summary_14_meta_90) %>%
  bind_rows(sim_summary_14_meta_180) %>%
  bind_rows(sim_summary_14_meta_110) %>%
  bind_rows(sim_summary_14_meta_120) %>%
  bind_rows(sim_summary_14_meta_130) %>%
  bind_rows(sim_summary_14_meta_150) %>%
  bind_rows(sim_summary_14_meta_220) %>%
  bind_rows(sim_summary_14_meta_270) %>%
  bind_rows(sim_summary_14_meta_365) %>%
  mutate(immunity_duration = 1/omega) %>%
  arrange(immunity_duration) %>%
  mutate(model = "meta",
         patches = 14)

write_csv(waning_results_14, file = "/Users/matthewhoyle/Github_R_projects/Hunter_Gatherer_models/Results/waning_results_14.csv")

waning_results_14

ggplot(waning_results_14, aes(immunity_duration, sum_persist)) +
  geom_line()+
  geom_point()+
  theme_bw()

Combined Results

Results from sinale and metapopulation models were comined into one data frame and visualised.

combined_waning <- waning_results %>%
  bind_rows(waning_results_single) %>%
  bind_rows(waning_results_7) %>%
  bind_rows(waning_results_3) %>%
  bind_rows(waning_results_14)

head(combined_waning)


#write_csv(combined_waning, file = "/Users/matthewhoyle/Github_R_projects/Hunter_Gatherer_models/Results/combined_waning_results.csv")

combined_plot <- ggplot(combined_waning, aes(immunity_duration, sum_persist, colour = as.factor(patches)))+
  geom_line(alpha=0.7, size=1)+
  geom_point(alpha=0.5, size=2)+
  geom_segment(x = -Inf, y = 50, xend = 141.5, yend = 50, linetype = "dashed", colour = "grey") +
  geom_segment(x = 5, y = 50, xend = 5, yend = -Inf, linetype = "dashed", colour = "grey") +
  geom_segment(x = 42.5, y = 50, xend = 42.5, yend = -Inf, linetype = "dashed", colour = "grey") +
  geom_segment(x = 91.5, y = 50, xend = 91.5, yend = -Inf, linetype = "dashed", colour = "grey") +
  geom_segment(x = 141.5, y = 50, xend = 141.5, yend = -Inf, linetype = "dashed", colour = "grey") +
  scale_y_continuous(breaks = seq(0, 100, 10)) +
  scale_x_continuous(breaks = seq(0, 360, 50)) +
  labs(x = "Duration of immunity (days)",
       y = "Probability of persistence after 3 years (%)", 
       colour = "No. Patches")+
  scale_color_discrete(type = wes_palettes$Zissou1[2:5],
                       labels = c("1", "3", "7", "14"))+
  theme_bw()

combined_plot

ggsave(filename = "combined_plot_patches.pdf", plot = combined_plot, device = "pdf", width = 7, height = 5, path = "/Users/matthewhoyle/Github_R_projects/Plots/Hunter_Gatherer_models")

ggplot(combined_waning, aes(immunity_duration, mean_percent_infected, colour = as.factor(patches)))+
  geom_line()+
  geom_point() +
  labs(x = "Duration of immunity",
       y = "Proportion infected at endpoint (%)", 
       colour = "Patches")+
  scale_color_discrete(type = wes_palettes$Zissou1[2:5],
                       labels = c("1", "3", "7", "14"))+
  theme_bw()

References